Expand Configuration and Management Manual (H06.03+)

ManualsBrandsHP ManualsServerHP Integrity NonStop H-Series

Expand Configuration

and Management

Manual

Abstract

This manual describes how to plan, configure, manage, and troubleshoot the Expand

subsystem on an HP Integrity NonStop™ NS-series server. The Expand subsystem

can connect as many as 255 geographically dispersed HP servers to create a network

with the reliability, capacity to preserve data integrity, and potential for expansion of a

single HP server. This manual includes detailed descriptions of SCF commands and

modifiers used with the Expand subsystem.

Product Version

Expand H06

Supported Release Version Updates (RVUs)

This manual supports H06.03 and all subsequent H-series RVUs until otherwise

indicated by its replacement publication.

Part Number Published

529522-002 July 2005

Summary of content (642 pages)

PAGE 1
Expand Configuration and Management Manual Abstract This manual describes how to plan, configure, manage, and troubleshoot the Expand subsystem on an HP Integrity NonStop™ NS-series server. The Expand subsystem can connect as many as 255 geographically dispersed HP servers to create a network with the reliability, capacity to preserve data integrity, and potential for expansion of a single HP server. This manual includes detailed descriptions of SCF commands and modifiers used with the Expand subsystem.
PAGE 2
Document History Part Number Product Version 529522-002 Expand H06 Published July 2005
PAGE 3
Expand Configuration and Management Manual Glossary Index What’s New in This Manual xxvii Manual Information xxvii New and Changed Information Examples Figures Tables xxvii About This Manual xxix Who Should Use This Manual xxix How This Manual Is Organized xxix Related Documents and Online Tools Notation Conventions xxxvii Abbreviations xli xxxiii Part I. Getting Started 1.
PAGE 4
2. Expand Overview Contents 2. Expand Overview Network Transparency 2-1 Interactive Access 2-1 Programmatic Access 2-2 Expand Subsystem and the NonStop Kernel 2-2 Multiple Communications Environments 2-5 Leased and Satellite Connections 2-5 X.
PAGE 5
3. Planning a Network Design (continued) Contents 3. Planning a Network Design (continued) When to Use a Single-Line Expand Line-Handler Process When to Use a Multi-Line Path 3-7 When to Use a Multi-CPU Path 3-9 Selecting Special Features 3-11 Multipacket Frame Feature 3-11 Variable Packet Size Feature 3-12 Congestion Control Feature 3-12 Designing the Network Topology 3-12 Common Network Topologies 3-12 Topology Limitations 3-14 Creating a Network Diagram 3-15 3-7 4.
PAGE 6
7. Configuring Direct-Connect and SatelliteConnect Lines Contents 7.
PAGE 7
Contents 8. Configuring Expand-Over-IP Lines (continued) 8.
PAGE 8
9. Configuring Expand-Over-ATM Lines (continued) Contents 9. Configuring Expand-Over-ATM Lines (continued) Step 2: Create a Profile for the Line-Handler Process ADD Profile Command 9-11 Example 9-11 Step 3: Create the Line-Handler Process 9-12 ADD DEVICE Command 9-12 Considerations 9-16 Examples 9-16 Step 4: Start the Line-Handler Process 9-17 Step 5: Start the Line 9-17 Profile Modifiers 9-17 Recommended Modifiers 9-18 Modifiers for Special Features 9-19 PEXQSATM Modifiers 9-19 9-10 10.
PAGE 9
10. Configuring Expand-Over-X.25 Lines (continued) Contents 10. Configuring Expand-Over-X.25 Lines (continued) Step 5: Start the Expand-Over-X.25 Line-Handler Process Step 6: Start the Expand-Over-X.25 Line 10-13 Profile Modifiers 10-13 Recommended Modifiers 10-13 Modifiers for Special Features 10-14 X25AM Line-Handler Process Modifiers 10-14 PEXQSNAM Modifiers 10-14 10-12 11.
PAGE 10
11. Configuring Expand-Over-SNA Lines (continued) Contents 11. Configuring Expand-Over-SNA Lines (continued) Profile Modifiers 11-16 Recommended Modifiers 11-16 Modifiers for Special Features 11-16 PEXQSNAM Modifiers 11-17 12.
PAGE 11
13. Configuring Multi-Line Paths Contents 13.
PAGE 12
Contents 14. Subsystem Control Facility (SCF) Commands (continued) 14.
PAGE 13
Contents 14. Subsystem Control Facility (SCF) Commands (continued) 14.
PAGE 14
. Subsystem Control Facility (SCF) Commands (continued) Contents 14. Subsystem Control Facility (SCF) Commands (continued) VERSION Command 14-111 Considerations 14-111 Examples 14-111 VERSION PROCESS Command 14-112 15.
PAGE 15
16. Expand Modifiers Contents Part IV. Reference Information 16.
PAGE 16
. Expand Modifiers (continued) Contents 16.
PAGE 17
. Subsystem Description (continued) Contents 17.
PAGE 18
17. Subsystem Description (continued) Contents 17.
PAGE 19
18. Managing the Network (continued) Contents 18.
PAGE 20
19. Tuning (continued) Contents 19. Tuning (continued) NAM Interface 19-12 Data Compression 19-12 Multi-Line Paths 19-13 Multi-CPU Paths 19-14 Network Topology 19-20 Summary of Tuning Strategies 19-21 Measuring and Mapping an Expand Network 19-22 What the Utilities Show 19-22 Using Measure 19-23 Measuring Passthrough Traffic 19-28 Setting Measurement Intervals 19-28 Tuning Examples 19-28 Example 1: Changing Packet Size 19-28 Example 2: Reducing Passthrough Traffic 19-31 20.
PAGE 21
20. Troubleshooting (continued) Contents 20. Troubleshooting (continued) Resolving Common Network Problems 20-32 Slow Response Time 20-32 Network Congestion 20-34 Node Not Available 20-34 Adding Low-Speed Lines to a Multi-Line Path Duplicate Node 20-37 20-37 A.
PAGE 22
Glossary Contents Glossary Index Examples Example 1-1. Example 1-2. Example 1-3. Example 6-1. Example 7-1. Example 8-1. Example 8-2. Example 8-3. Example 8-4. Example 8-5. Example 8-6. Example 8-7. Example 8-8. Example 8-9. Example 8-10. Example 9-1. Example 9-2. Example 9-3. Example 9-4. Example 9-5. Example 14-1. Example 14-2. Example 14-3. Example 14-4. Example 14-5. Example 14-6. Example 14-7.
PAGE 23
Examples (continued) Contents Examples (continued) Example 14-8. Example 14-9. Example 14-10. Example 14-11. Example 14-12. Example 14-13. Example 14-14. Example 14-15. Example 14-16. Example 14-17. Example 14-18. Example 14-19. Example 14-20. Example 14-21. Example 14-22. Example 14-23. Example 14-24. Example 14-25. Example 14-26. Example 14-27. Example 14-28. Example 14-29. Example 14-30. Example 14-31. Example 14-32. Example 14-33. Example 14-34. Example 14-35. Example 19-1. Example 19-2. Example 19-3.
PAGE 24
Examples (continued) Contents Examples (continued) Example 19-4. Example 19-5. Example 19-6. Example 19-7. Example 19-8. Example 20-1. Example 20-2. Example 20-3. Example 20-4. Example 20-5. Example 20-6. Example 20-7. Example 20-8. Example 20-9. Example 20-10. Example 20-11. Example 20-12. Example 20-13.
PAGE 25
Figures (continued) Contents Figures (continued) Figure 9-1. Figure 9-2. Figure 9-3. Figure 10-1. Figure 10-2. Figure 11-1. Figure 11-2. Figure 11-3. Figure 12-1. Figure 12-2. Figure 13-1. Figure 13-2. Figure 14-1. Figure 15-1. Figure 17-1. Figure 17-2. Figure 17-3. Figure 17-4. Figure 17-5. Figure 17-6. Figure 17-7. Figure 17-8. Figure 17-9. Figure 17-10. Figure 17-11. Figure 17-12. Figure 17-13. Figure 17-14. Figure 17-15. Figure 17-16. Figure 17-17. Figure 17-18. Figure 17-19. Figure 17-20.
PAGE 26
Figures (continued) Contents Figures (continued) Figure 19-2. Figure 19-3. Figure 19-4. Figure 19-5. Figure 19-6. Figure 20-1. Application Data Flow for Expand-Over-IP 19-7 CPU Matching 19-17 Pair Count Balancing for Neighbors and Non-Neighbors Passthrough Traffic 19-20 Packet Size/Bandwidth Comparison 19-31 Network Problem Hierarchy 20-3 19-18 Tables Table i. Table ii. Table iii. Table iv. Table v. Table vi. Table 1-1. Table 1-2. Table 1-3. Table 1-4. Table 1-5. Table 1-6. Table 1-7. Table 1-8.
PAGE 27
Tables (continued) Contents Tables (continued) Table 13-5. Table 13-6. Table 13-7. Table 14-1. Table 14-2. Table 14-3. Table 14-4. Table 14-5. Table 14-6. Table 14-7. Table 14-8. Table 14-9. Table 14-10. Table 15-1. Table 15-2. Table 15-3. Table 16-1. Table 16-2. Table 16-3. Table 16-4. Table 18-1. Table 18-2. Table 18-3. Table 18-4. Table 18-5. Table 18-6. Table 18-7. Table 18-8. Table 18-9. Table 18-10. Table 18-11. Table 18-12. Table 19-1. Table 19-2. Table 19-3.
PAGE 28
Tables (continued) Contents Tables (continued) Table 20-1. Table 20-2. Table 20-3. Table 20-4. Table 20-5. Table 20-6. Table 20-7. Table 20-8. Table 20-9. Table 20-10. Table 20-11. Table 20-12. Table 20-13. Table 20-14. Table 20-15. Table 20-16. Table 20-17. Table B-1.
PAGE 29
What’s New in This Manual Manual Information Expand Configuration and Management Manual Abstract This manual describes how to plan, configure, manage, and troubleshoot the Expand subsystem on an HP Integrity NonStop™ NS-series server. The Expand subsystem can connect as many as 255 geographically dispersed HP servers to create a network with the reliability, capacity to preserve data integrity, and potential for expansion of a single HP server.
PAGE 30
What’s New in This Manual New and Changed Information Expand Configuration and Management Manual—529522-002 xxviii
PAGE 31
About This Manual The Expand Configuration and Management Manual describes how to plan, configure, and manage the Expand subsystem on an HP Integrity NonStop™ NS-series server.
PAGE 32
Part I Contents About This Manual Part I Contents Part I, Getting Started, consists of Sections 1 through 4. Table i summarizes the contents of Part I. Table i. Summary of Contents—Part I Section Title Contents 1 Configuration Quick Start Provides the basic information required to enable you to quickly define, start, and modify Expand linehandler process. 2 Expand Overview Describes the Expand subsystem’s major features and capabilities.
PAGE 33
Part III Contents About This Manual Table ii. Summary of Contents—Part II (page 2 of 2) Section Title Contents 11 Configuring ExpandOver-SNA Lines Explains how to configure and start single-line Expand-over-SNA line-handler processes. 12 Configuring ExpandOver-ServerNet Lines Explains how to configure Expand-over-ServerNet line-handler processes. 13 Configuring Multi-Line Paths Explains how to configure and start multi-line paths.
PAGE 34
Part V Contents About This Manual Part V Contents Part V, Management, Tuning, and Troubleshooting, consists of Sections 18 through 20. Table v summarizes the contents of Part V. Table v. Summary of Contents—Part IV Section Title Contents 18 Managing the Network This section explains how to access network resources, set up network security, and monitor, reconfigure, and control an Expand network.
PAGE 35
Related Documents and Online Tools About This Manual Related Documents and Online Tools The Integrity NonStop NS-series server manual set contains manuals that describe how to configure both an entire system and individual hardware and software components, such as peripheral devices and communications software.
PAGE 36
SCF Subsystem Configuration Manuals About This Manual • SCF Reference Manual for H-Series RVUs This manual describes the operation of the Subsystem Control Facility (SCF) on H-series RVUs and how it is used to configure, control, and inquire about supported SCF subsystems. SCF is the configuration and management tool used by persons responsible for configuring system objects or monitoring their status.
PAGE 37
Related HP Manuals About This Manual Related HP Manuals You may need to refer to the following HP manuals when configuring and managing an Expand network: • ASAP Migration Guide for NSX and OMF Users This guide introduces the Availability Statistics and Performance (ASAP) product to users of the Network Statistics Extended (NSX) and Object Monitoring Facility (OMF) products.
PAGE 38
Guided Procedure for Configuring a ServerNet Node About This Manual • ServerNet Cluster 6780 Planning and Installation Guide This manual describes the installation and planning for the 6780 ServerNet Cluster switch. • SNAX/XF and SNAX/APN Configuration and Management Manual This manual describes how to configure the SNAX/XF and SNAX/APN communications subsystems. It includes detailed descriptions of the Subsystem Control Facility (SCF) commands used with the SNAX/XF and SNAX/APN subsystems.
PAGE 39
Notation Conventions About This Manual Notation Conventions Hypertext Links Blue underline is used to indicate a hypertext link within text. By clicking a passage of text with a blue underline, you are taken to the location described. For example: This requirement is described under Backup DAM Volumes and Physical Disk Drives on page 3-2. General Syntax Notation The following list summarizes the notation conventions for syntax presentation in this manual. UPPERCASE LETTERS.
PAGE 40
General Syntax Notation About This Manual each side of the list, or horizontally, enclosed in a pair of brackets and separated by vertical lines. For example: LIGHTS [ ON ] [ OFF ] [ SMOOTH [ num ] ] K [ X | D ] address-1 { } Braces. A group of items enclosed in braces is a list from which you are required to choose one item. The items in the list may be arranged either vertically, with aligned braces on each side of the list, or horizontally, enclosed in a pair of braces and separated by vertical lines.
PAGE 41
Notation for Messages About This Manual Line Spacing. If the syntax of a command is too long to fit on a single line, each continuation line is indented three spaces and is separated from the preceding line by a blank line. This spacing distinguishes items in a continuation line from items in a vertical list of selections. For example: ALTER [ / OUT file-spec / ] CONTROLLER [ , attribute-spec ]...
PAGE 42
Notation for Subnet About This Manual either vertically, with aligned braces on each side of the list, or horizontally, enclosed in a pair of braces and separated by vertical lines. For example: LBU { X | Y } POWER FAIL process-name State changed from old-objstate to objstate { Operator Request. } { Unknown. } | Vertical Line. A vertical line separates alternatives in a horizontal list that is enclosed in brackets or braces. For example: Transfer status: { OK | Failed } % Percent Sign.
PAGE 43
Abbreviations About This Manual Abbreviations The following list defines abbreviations and acronyms used in this guide. Both industrystandard terms and HP terms are included. API. Application Program Interface ATM. Asynchronous Transfer Mode ATM3SA. ATM 3 ServerNet Adapter ASAP. Availability Statistics and Performance CAP. Communications Access Protocol CLIP. Communications Line Interface Processor ConMgr. Concentrator Manager Process DLC. Data Link Control DSM. Distributed Systems Management DV.
PAGE 44
Abbreviations About This Manual IP. Internet Protocol LAN. Local Area Network LNP. Logical Network Partitioning LU. Logical Unit MPT. Multiple Path Table MSH. Modified Split Horizon NAM. Network Access Method NCP. Network Control Process NRT. Network Routing Table OOS. Out Of Sequence OSI. Open Systems Interconnection OSS. Open System Services PIN. Process Identification Number PU. Physical Unit PVC. Permanent Virtual Circuit RPT. Reverse Pairing Table SAN. System Area Network SCF.
PAGE 45
Abbreviations About This Manual TCP/IP. Transmission Control Protocol/Internet Protocol TF. Time Factor TFTP. Trivial File Transfer Protocol UDP. User Datagram Protocol WAN. Wide Area Network X25AM. X.25 Access Method $NCP. Network Control Process name $ZEXP. Expand Manager Process name $ZNET. Subsystem Control Point process name $ZNUP. Network Utility Process name $ZPM. Persistence Manager Process name $ZZKRN. Kernel Subsystem Manager Process name $ZZLAN. SLSA Subsystem Manager Process name $ZZSCL.
PAGE 46
Abbreviations About This Manual Expand Configuration and Management Manual—529522-002 xliv
PAGE 47
Part I.
PAGE 48
Part I.
PAGE 49
1 Configuration Quick Start This section provides the basic information required to enable you to quickly and easily define and start an Expand line-handler process. This procedure requires that you use the default values provided by the Expand subsystem for most configuration modifiers. If you want a customized configuration, or if you want to change your configuration, you must refer to the sections provided in Part II, Configuring the Expand Subsystem.
PAGE 50
Configuration Quick Start Task 1: Configure and Start $NCP Task 1: Configure and Start $NCP The network control process ($NCP) is responsible for initiating and terminating serverto-server connections and maintaining network-related system tables, including routing information. $NCP must be running at every node in the Expand network before Expand lines can be started. To configure and start the network control process, perform the following steps: 1.
PAGE 51
Configuration Quick Start Creating a Persistent Version of the Expand Manager Process 2. You can also start the Expand manager process at system startup by including the following command in the system startup file: OZEXP / NAME $ZEXP, OUT $ZHOME, PRI 180, NOWAIT, & CPU primary / backup 3.
PAGE 52
Task 3: Add the Expand Line-Handler Profile(s) Configuration Quick Start Task 3: Add the Expand Line-Handler Profile(s) HP provides profiles, which contain modifiers and default modifier values, for each type of Expand line-handler process. You can use these profiles to create profiles for your Expand line-handler processes. To add a profile for an Expand line handler, perform the following steps: Note.
PAGE 53
Where to Find More Information About This Task Configuration Quick Start Table 1-2. Profiles for Line-Logical Devices Profile Name Type of Line-Logical Device PEXQMSWN Direct-connect PEXQMSAT Satellite-connect PEXQMNAM Expand-over-NAM PEXQMATM Expand-over-ATM PEXQMIP Expand-over-IP The following rules apply when creating profiles for lines in a multi-line path: • • • You can configure a maximum of eight lines in a multi-line path.
PAGE 54
Configuration Quick Start Task 4: Add the Expand Line-Handler Process Task 4: Add the Expand Line-Handler Process The Expand subsystem supports a variety of different protocols and communications methods to enable you to connect systems together in local area network (LAN) and wide area network (WAN) topologies. The following types of Expand line-handler processes can be configured: • • • • • • • Direct-connect Satellite-connect Expand-over-IP Expand-over-ATM Expand-over-X.
PAGE 55
Creating a Single-Line Expand Line-Handler Process Configuration Quick Start Example 1-1. SCF STATUS ADAPTER Command WAN Manager STATUS ADAPTER for ADAPTER State........... STARTED \NODEA.$ZZWAN.#S01 Number of clips. 3 Clip 1 status : CONFIGURED Clip 2 status : CONFIGURED Clip 3 status : CONFIGURED WAN Manager STATUS SERVER for CLIP State :......... STARTED \NODEA.$ZZWAN.#S01.1 Path A..........: CONFIGURED Path B..........: CONFIGURED Number of lines. 2 Line............ 0 Line............
PAGE 56
Creating a Single-Line Expand Line-Handler Process Configuration Quick Start 3. Using the name of the SWAN concentrator with the available WAN line from Step 2a, determine the names of the preferred and alternate NonStop TCP/IP processes configured for the SWAN concentrator. -> INFO ADAPTER $ZZWAN.
PAGE 57
Creating a Single-Line Expand Line-Handler Process Configuration Quick Start Example 1-3. SCF STATUS PROCESS Command -> STATUS PROCESS $ZB018 TCPIP Status process \NODEA.$ZB018 Status: Started PPID............. ( 0,319) BPID................ ( 1,292) Proto TCP TCP TCP Faddr 0.0.0.0 0.0.0.0 0.0.0.0 Status LISTEN LISTEN LISTEN Laddr 0.0.0.0 0.0.0.0 0.0.0.0 Lport ftp finger echo Fport * * * SendQ 0 0 0 RecvQ 0 0 0 -> STATUS PROCESS $ZB01C TCPIP Status process \NODEA.$ZB01C Status: Started PPID......
PAGE 58
Creating a Single-Line Expand Line-Handler Process Configuration Quick Start Table 1-3. SCF ADD DEVICE Command Worksheet Parameter Value/Description device_name The name you want to assign to the Expand line-handler process. name The name of the profile you created in Task 3: Add the Expand LineHandler Profile(s) on page 1-4.
PAGE 59
Creating a Single-Line Expand Line-Handler Process Configuration Quick Start If you want to use IPv6 communications, add the device as follows: -> ADD DEVICE $ZZWAN.#device_name, PROFILE name,& IOPOBJECT $SYSTEM.SYS00.LHOBJ, CPU cpunum, ALTCPU altcpu,& TYPE (63,0), RSIZE 0, PATHTF 2, NEXTSYS sysnum,& ASSOCIATEDEV tcp6sam_process, IPVER_IPV6,& V6SRCIPADDR ipv6srcaddress, V6DESTIPADDR ipv6destaddress,& SRCIPPORT sipport, DESTIPPORT dipport Note.
PAGE 60
Creating a Single-Line Expand Line-Handler Process Configuration Quick Start Table 1-4. SCF ADD DEVICE Syntax: Expand-Over-IP (page 2 of 2) Parameter Description sipaddr The IP address used by the NonStop TCP/IP process specified by tcpip_process. The address must be specified by number. sipport The UDP port number used by this Expand-over-IP line-handler process. Valid values are in the range 0 through 65536. Do not use well-known ports in the range 0 through 1023.
PAGE 61
Creating a Single-Line Expand Line-Handler Process Configuration Quick Start Use the following command syntax for an Expand-over-ATM line-handler process using an ATMSAP connection through the SLSA subsystem: -> ADD DEVICE $ZZWAN.#device_name, PROFILE name,& IOPOBJECT $SYSTEM.SYS00.LHOBJ, CPU cpunum, ALTCPU altcpu,& TYPE (63,0), RSIZE 0, PATHTF 3, CALLTYPE_ATMSAP,& LIFNAME lif-name, NEXTSYS sysnum Note.
PAGE 62
Creating a Single-Line Expand Line-Handler Process Configuration Quick Start Expand-Over-X.25, Expand-Over-SNA and Expand-OverServerNet Line-Handler Processes Expand-over-X.25, Expand-over-SNA, and Expand-over-ServerNet line-handler processes require the services of the following software components: • • • The Expand-over-X.25 line-handler process must be associated with an X25AM process. It uses a NAM subdevice defined for the X25AM process.
PAGE 63
Creating a Multi-Line Path Configuration Quick Start Table 1-6. SCF ADD DEVICE Syntax: Expand-Over-X.25, Expand-Over-SNA, and Expand-Over-ServerNet (page 2 of 2) Parameter process Description • • subdevice • • • • sysnum For Expand-over-X.25: the name of an X25AM line-handler process. For Expand-over-SNA: the name of a SNAX/APN line-handler process. For Expand-over-ServerNet: must be $ZZSCL. For Expand-over-X.25: the name of an X25AM subdevice defined for the X25AM process specified by process.
PAGE 64
Where to Find More Information About This Task Configuration Quick Start 2. To create lines in the multi-line path (called line-logical devices), use the SCF ADD DEVICE command syntax shown for configuring single-line Expand line-handler processes (see Creating a Single-Line Expand Line-Handler Process on page 1-6), but with the following exceptions: • Use the TYPE modifier values as shown in Table 1-8. Table 1-8.
PAGE 65
Configuration Quick Start Task 5: Start the Expand Line-Handler Process Section 12, Configuring Expand-Over-ServerNet Lines Section 13, Configuring Multi-Line Paths Task 5: Start the Expand Line-Handler Process Start the single-line Expand line-handler process or path-logical device. When you use this command to start a path-logical device, the line-logical devices associated with the path are also started. -> START DEVICE $ZZWAN.
PAGE 66
Configuration Quick Start Starting Lines in a Multi-Line Path Expand Configuration and Management Manual—529522-002 1-18
PAGE 67
2 Expand Overview The Expand subsystem enables you to connect as many as 255 geographically dispersed NonStop servers to create a network with the reliability, capacity to preserve data integrity, and potential for expansion of a single NonStop server.
PAGE 68
Programmatic Access Expand Overview Programmatic Access When accessing a file or another resource programmatically across an Expand network, you use the same procedure calls you would use when writing a local application. With a few exceptions, applications that were written to run in a local environment can be used virtually unchanged in a network environment. Expand Subsystem and the NonStop Kernel The Expand subsystem is an extension of the HP NonStop Kernel operating system.
PAGE 69
Expand Subsystem and the NonStop Kernel Expand Overview Single-Server Process Communications Figure 2-1 illustrates how a process on one processor uses the file system to make an inquiry of a process residing on another processor in the same server. The message system relays the request through the ServerNet system area network (ServerNet SAN). Figure 2-1.
PAGE 70
Expand Subsystem and the NonStop Kernel Expand Overview Multi-Node Process Communications Figure 2-2 illustrates the same file-system request as Figure 2-1, except that the disk process resides on another node in the network rather than on another processor in the same server. Figure 2-2.
PAGE 71
Multiple Communications Environments Expand Overview Multiple Communications Environments Nodes in an Expand network can be connected using a variety of data communications technologies and protocols. A single network can consist of any combination of these different data communications methods. Nodes in an Expand network can be connected by • • • • • • Full-duplex leased lines or satellite connections using the High-Level Data Link Control (HDLC) protocol X.
PAGE 72
Systems Network Architecture (SNA) Networks Expand Overview Systems Network Architecture (SNA) Networks SNA was developed by IBM for connecting IBM systems and networks. Expand-overSNA connections are provided with the HP SNAX/Advanced Peer Networking (SNAX/APN) product. The Expand subsystem uses the NETNAM protocol to communicate with the SNAX/APN line-handler process. Internet Protocol (IP) Networks An IP network adheres to the Internet Protocol—a computer-industry standard protocol.
PAGE 73
Distributed Control Expand Overview Distributed Control The control function of the Expand subsystem is distributed throughout the network. Unlike a hierarchical network, in which a central computer, or host, controls the communications environment, nodes in an Expand network communicate with each other as peers. Distributed networks have the following additional advantages: • • • Distributed applications. Applications can be distributed so that multiple nodes share the processing load.
PAGE 74
Priority Routing Expand Overview Priority Routing You can assign different priorities to messages sent over an Expand network. Priority routing allows an important message to reach its destination even when the network is congested. Fault-Tolerant Operation Using careful configuration and network-topology design, you can configure an Expand network to be continuously available. You can configure as many as eight lines between the same two nodes using the Expand subsystem’s multi-line path feature.
PAGE 75
Subsystem Control Facility (SCF) Expand Overview Note. Refer to Part III of this manual, Subsystem Control Facility (SCF), for information on managing Expand using SCF. Subsystem Control Facility (SCF) SCF is a Distributed Systems Management (DSM) interface that can be used interactively to control, configure, and monitor the Expand subsystem. The SCF interfaces to the Expand and wide area network (WAN) subsystems are used to configure and manage the Expand subsystem.
PAGE 76
Online Expansion and Reconfiguration Expand Overview Online Expansion and Reconfiguration You can add a new node or new lines to a network or move an existing node to a different location without disrupting network activity. You can make changes to your Expand configuration online using the Subsystem Control Facility (SCF) interfaces to the Expand and WAN subsystems. Table 2-1 shows the online expansion and reconfiguration tasks that can be performed with these interfaces. Table 2-1.
PAGE 77
Enhanced Security Techniques Expand Overview Enhanced Security Techniques The Safeguard security system enhances the security provided by both the Expand subsystem and the NonStop Kernel operating system. Safeguard enables you to set password expiration dates, create access control lists, and audit file access. For an even greater level of security, data encryption devices are available from the HP Atalla subsidiary.
PAGE 78
Expand Overview Enhanced Security Techniques Expand Configuration and Management Manual—529522-002 2- 12
PAGE 79
3 Planning a Network Design This section describes the network design decisions you must make before installing and configuring a new Expand network or when modifying an existing Expand network. Topics described in this section include • • • • • Selecting Line Protocols on page 3-1 Defining Paths Between Systems on page 3-6 Selecting Special Features on page 3-11 Designing the Network Topology on page 3-12 Creating a Network Diagram on page 3-15 Note.
PAGE 80
Planning a Network Design Satellite Connections Satellite Connections The satellite-connect line-handler process implements the satellite-efficient version of the HDLC protocol, HDLC Extended Mode. HDLC Extended Mode allows a maximum window size of 61 frames (the maximum window size is the number of outstanding frames that can be sent before an acknowledgment is required) and implements the selective-reject feature. Selective reject causes only frames that arrive in error to be retransmitted.
PAGE 81
Planning a Network Design • • Systems Network Architecture (SNA) Connections Low capital cost/high connectivity. X.25 provides a way to connect a large number of systems through a single line between a NonStop server and an X.25 network. This feature can lower communications capital costs by reducing the number of modems and controller ports that must be purchased. For example, a fully connected network of 4 servers requires 6 links, 12 modems, and 12 hardware ports. An X.
PAGE 82
Planning a Network Design Internet Protocol (IP) Networks Internet Protocol (IP) Networks The IP suite is an important industry standard. Expand-over-IP allows NonStop systems to be interconnected via inexpensive IP-based routers, making a separate Expand network unnecessary. Expand-over-IP uses a NonStop TCP/IP process to implement the TCP/IP protocol stack. The Expand-over-IP line-handler process communicates with the NonStop TCP/IP process through the shared memory of the QIO subsystem.
PAGE 83
Planning a Network Design Asynchronous Transfer Mode (ATM) Networks When you are planning your Expand-over-IP environment, you may use LNP to control over which network interfaces (IP addresses) the Expand line-handler processes run. See Step 1 (B): Select a Process and SUBNET for NonStop TCP/IPv6 Use on page 8-11 for examples of working with logical network partitioning. To determine which TCP/IP subsystem is running on your system, use the SCF LISTDEV TCPIP command. The text after the last period (.
PAGE 84
Planning a Network Design • ServerNet Connections Passthrough capability. Packets sent over an ATM network path can be forwarded to another Expand line-handler process, which can be a different line type and in a different processor. ServerNet Connections The Expand-over-ServerNet line-handler process provides connectivity to a ServerNet cluster, which uses this process to provide a very high speed proprietary interconnect between systems over a limited geographic range.
PAGE 85
Planning a Network Design • • When to Use a Single-Line Expand Line-Handler Process Enable the variable packet size feature Enable the congestion control feature When to Use a Single-Line Expand Line-Handler Process Single-line Expand line-handler processes are less expensive and require somewhat less processing time than multi-line paths. However, they lack the fault-tolerance that multi-line paths and multi-CPU paths provide.
PAGE 86
When to Use a Multi-Line Path Planning a Network Design Figure 3-1. Multi-Line Path With Eight Lines and Two SWAN Concentrators $LINE1 $LINE2 SWAN $LINE3 $LINE4 $PATH1 $LINE5 $LINE6 SWAN $LINE7 $LINE8 CDT 025.
PAGE 87
When to Use a Multi-CPU Path Planning a Network Design Figure 3-2 illustrates an eight-line configuration. Figure 3-2. Multi-Line Path With Eight Lines and Eight SWAN Concentrators $LINE1 SWAN $LINE2 SWAN $LINE3 SWAN $LINE4 SWAN $LINE5 SWAN $LINE6 SWAN $LINE7 $X25AM1 SWAN $LINE8 $X25AM2 SWAN $PATH2 CDT 026.CDD When to Use a Multi-CPU Path The Expand multi-CPU feature enables you to connect multiple Expand line-handler processes, each in a separate processor, between two nodes.
PAGE 88
When to Use a Multi-CPU Path Planning a Network Design • Maximum throughput is significantly increased, especially for Expand-over-IP connections. An Expand-over-IP line-handler process and its associated NonStop TCP/IP process must be configured in the same processor pair, placing the burden of processing the entire communications protocol stack for each Expand-over-IP line on one processor.
PAGE 89
Planning a Network Design Selecting Special Features For more information about multi-CPU paths, refer to Multi-CPU Feature on page 17-72. Note. You use the SUPERPATH_ON modifier to configure an Expand line-handler process as part of a multi-CPU path. If you configure parallel paths between two nodes without using the SUPERPATH_ON modifier, only one path is used at a given time.
PAGE 90
Planning a Network Design Variable Packet Size Feature Variable Packet Size Feature The variable packet size feature is a performance enhancement designed to increase bulk transfers over all connection types.
PAGE 91
Common Network Topologies Planning a Network Design • • Mesh Mixed The star, tree, ring, bus, and mesh topologies are illustrated in Figure 3-4. A mixed topology is a combination of more than one type of topology. The split-star and tri-star topologies are extensions of the star topology. Figure 3-4. Common Network Topologies STAR TREE RING BUS MESH VST028 Star Topology In a star topology, all systems join at a central node, creating a star-shaped configuration.
PAGE 92
Planning a Network Design Topology Limitations Split-Star Topology Used for ServerNet clusters, the split-star topology connects two star topologies. Each star contains a cluster switch. The two cluster switches are connected by fiber optic cables, each of which can be up to one kilometer in length. This topology can be used for more than nine and fewer than 16 nodes. For examples of this topology, refer to Section 4, Planning for ServerNet Clusters.
PAGE 93
Planning a Network Design Creating a Network Diagram restriction limits the size of any network configured as a fully connected mesh to 255 nodes. Creating a Network Diagram Before you configure your Expand network, HP recommends that you create a diagram of the complete network topology. This network diagram shows the network nodes and the lines that connect the nodes. This type of diagram can help you and the operations staff monitor systems, recognize problems, and prepare for configuration changes.
PAGE 94
Creating a Network Diagram Planning a Network Design Figure 3-5. Network Diagram \LA Node1 \DALLAS Node 2 $PATH1 $PATH1 $LINE1 $LINE1 $LINE2 $LINE2 $LINE3 $LINE3 SWAN SWAN $LINEA $LINEB SWAN SWAN SWAN SWAN $LINEA $LINEB SWAN $LINEC \BOISE Node 3 SWAN $LINEC \CHICAGO Node 4 CDT 030.
PAGE 95
4 Planning for ServerNet Clusters This section describes how to plan for the configuration of Expand over ServerNet clusters, discusses considerations for ServerNet topologies, and provides examples of configuring Expand over ServerNet clusters, ServerNet clusters in combination with ServerNet/FX, ServerNet clusters in combination with ATM and IP networks, and ServerNet clusters with other communication methods. You can configure Expand over ServerNet clusters by using either OSM or SCF.
PAGE 96
Planning for ServerNet Clusters • • • • • • Configuration Considerations for Expand and ServerNet Clusters Each system in an Expand network can support up to 255 Expand line-handler processes. A node can only belong to one ServerNet cluster. The Expand manager process, $ZEXP, must be configured and started. The NonStop ServerNet cluster monitor process, $ZZSCL, must be configured and started in all Integrity NonStop nodes connected to a ServerNet cluster. As of G06.
PAGE 97
Planning for ServerNet Clusters ServerNet Clusters Coexisting With ATM or IP Networks ServerNet Clusters Coexisting With ATM or IP Networks Interoperability is supported between a ServerNet cluster and an ATM network using ATM adapters or between a ServerNet cluster and an IP network using Ethernet, Fast Ethernet, Gigabit Ethernet, or ATM adapters.
PAGE 98
Planning for ServerNet Clusters Examples of ServerNet Clusters Coexisting With ATM or IP Examples of ServerNet Clusters Coexisting With ATM or IP The following examples are provided in this subsection: • • • ServerNet Clusters Connected By ATM or IP Lines on page 4-5 ServerNet Clusters Connected By a Single ATM or IP Line (Not Recommended) on page 4-6 ServerNet Clusters Using Layered Topology With Connections to Nodes Outside the Cluster on page 4-7 Expand Configuration and Management Manual—529522-002
PAGE 99
Examples of ServerNet Clusters Coexisting With ATM or IP Planning for ServerNet Clusters ServerNet Clusters Connected By ATM or IP Lines All the systems in Figure 4-1 are Integrity NonStop NS-series systems. Groups of systems at separate locations can use clustering technology to increase performance within each location and connect to each other by using ATM or IP lines.
PAGE 100
Examples of ServerNet Clusters Coexisting With ATM or IP Planning for ServerNet Clusters ServerNet Clusters Connected By a Single ATM or IP Line (Not Recommended) All the systems in Figure 4-2 are Integrity NonStop NS-series systems. This topology is not recommended as a solution for enabling communication between two ServerNet clusters due to higher processor use, traffic bottlenecks, and a lack of overall network fault tolerance.
PAGE 101
Planning for ServerNet Clusters Examples of ServerNet Clusters Coexisting With ATM or IP ServerNet Clusters Using Layered Topology With Connections to Nodes Outside the Cluster In this example, all the systems directly connected to the cluster switches are Integrity NonStop NS-series systems. (See Figure 4-3.) This ServerNet cluster uses the layered topology (for more information about the layered topology, see the ServerNet Cluster 6780 Planning and Installation Guide ).
PAGE 102
Examples of ServerNet Clusters Coexisting With ATM or IP Planning for ServerNet Clusters Figure 4-3.
PAGE 103
Part II. Configuring the Expand Subsystem Part II consists of the following sections, which provide an overview of the configuration process and explain how to configure the various types of Expand line-handler processes: Section 5 Configuration Overview Section 6 Configuring the Network Control Process Section 7 Configuring Direct-Connect and Satellite-Connect Lines Section 8 Configuring Expand-Over-IP Lines Section 9 Configuring Expand-Over-ATM Lines Section 10 Configuring Expand-Over-X.
PAGE 104
Part II.
PAGE 105
5 Configuration Overview This section provides an overview of the Expand subsystem configuration process. Before using this section and the remaining sections in this manual, you should be familiar with the following: • • • • Section 3, Planning a Network Design. This section describes network design considerations such as selecting line protocols.
PAGE 106
Summary of Configuration Steps Configuration Overview Summary of Configuration Steps Configuring the Expand subsystem involves a number of steps. Table 5-1 lists each step and indicates where in this manual the step is described. Table 5-1. Configuration Steps Step Description Where This Step Is Described 1. Start the Expand manager process. For step 1, information is located in Starting the Expand Manager Process on page 5-4 of this section. 2.
PAGE 107
Creating a Profile Configuration Overview Creating a Profile A profile template is a disk file that contains modifiers and default modifier values. HP provides profile templates for the network control process ($NCP) and for the different types of Expand line-handler processes. Table 5-2 lists the profile templates for the Expand subsystem. These profile templates are installed in $SYSTEM.SYSnn. The modifiers in each profile template are described in the sections listed in Table 5-1.
PAGE 108
Configuration Overview Creating Wide Area Network (WAN) Subsystem Devices Creating Wide Area Network (WAN) Subsystem Devices The network control process and Expand line-handler processes are defined as WAN subsystem devices. The DEVICE object represents $NCP and Expand line-handler processes in the WAN subsystem. You use the WAN subsystem SCF ADD DEVICE command to create the network control process and Expand line-handler processes.
PAGE 109
6 Configuring the Network Control Process This section explains how to configure and start the network control process ($NCP). Configuring and starting $NCP involves the following steps: Step 1: Create a Profile for $NCP on page 6-1 Step 2: Create $NCP on page 6-2 Step 3: Start $NCP on page 6-4 You perform all these steps using the SCF interface to the WAN subsystem. This section also describes the $NCP profile modifiers in $NCP Modifiers on page 6-4.
PAGE 110
Configuring the Network Control Process Example modifier_value is the value you want to assign to the modifier specified by modifier_keyword. Specifying a value in modifier_value assigns a new value to modifier_keyword in profile_name. Default values and ranges of values for modifiers in the PEXPNCP profile are described in $NCP Modifiers on page 6-4. Example The following example creates a profile named NCPPROF1. The ALGORITHM modifier in the profile is set to 1 to specify split horizon.
PAGE 111
Configuring the Network Control Process Considerations CPU cpunum indicates the processor where $NCP will normally execute. HP recommends that you configure $NCP to run in processor 0. ALTCPU altcpunum indicates the processor where the backup $NCP will normally execute. HP recommends that you configure the backup $NCP to run in processor 1. TYPE (62,6) is the device type and subtype for $NCP. The device type is always 62 and the subtype is always 6 for $NCP.
PAGE 112
Step 3: Start $NCP Configuring the Network Control Process Step 3: Start $NCP To start $NCP, you use the WAN subsystem SCF START DEVICE command. The command syntax is as follows: START DEVICE $ZZWAN.#NCP To make sure $NCP has started successfully, enter the following command at the TACL prompt: > STATUS $NCP If $NCP was started successfully, you will see a display similar to Example 6-1: Example 6-1.
PAGE 113
Configuring the Network Control Process $NCP Modifiers ALGORITHM n Default: Units: Range: 0 (MSH) Not applicable 0 or 1 This modifier identifies $NCP routing algorithm to be used. Specify 0 for MSH or 1 for split horizon (SH). The ALGORITHM modifier must be set to the same value on all systems in the network. Modified split horizon (MSH) and split horizon (SH) algorithms are explained in detail in Routing Algorithms on page 17-27.
PAGE 114
Configuring the Network Control Process $NCP Modifiers FRAMESIZE n Default: Units: Range: 132 Words 64 through 250 The $NCP FRAMESIZE modifier specifies the maximum packet size that $NCP can send in the network. This value must be less than or equal to the Expand line-handler process’s FRAMESIZE modifier but, it cannot be greater than 250. It is not required that this modifier be the same for each $NCP in the network.
PAGE 115
7 Configuring Direct-Connect and Satellite-Connect Lines The direct-connect line-handler process implements the High-Level Data Link Control (HDLC) Normal protocol and operates with conventional voice-grade leased-line and switched-line facilities, private facilities, and fractional Transmission Group 1 (T1) facilities. The satellite-connect line-handler process implements the satellite-efficient version of the HDLC protocol, HDLC Extended mode.
PAGE 116
Configuring Direct-Connect and Satellite-Connect Lines Required Hardware and Software Required Hardware and Software Several hardware and software components are required in addition to the directconnect or satellite-connect line-handler process to provide direct-connect or satelliteconnect connectivity. These components are illustrated in Figure 7-1 and are explained in the following subsections. Figure 7-1.
PAGE 117
Configuring Direct-Connect and Satellite-Connect Lines QIO Subsystem QIO Subsystem QIO is a mechanism for transferring data between processes through a shared memory segment. The QIO subsystem is preconfigured and started during the system load sequence. The QIO subsystem must be started before Expand line-handler processes can be started. For information about the QIO subsystem, refer to the QIO Configuration and Management Manual.
PAGE 118
Configuring Direct-Connect and Satellite-Connect Lines Ethernet 4 ServerNet Adapter (E4SA) Ethernet 4 ServerNet Adapter (E4SA) The E4SA is a double-high ServerNet adapter that supports four Ethernet interfaces and communicates with multiple processors through dual ServerNet interfaces to the ServerNet fabrics. Two E4SAs are used to connect a SWAN concentrator to a processor. For information about E4SAs, refer to the LAN Configuration and Management Manual.
PAGE 119
Configuring Direct-Connect and Satellite-Connect Lines ServerNet Wide Area Network (SWAN) Concentrator ServerNet Wide Area Network (SWAN) Concentrator The SWAN concentrator is a communications device that provides WAN connections. HP recommends that you configure your satellite-connect or direct-connect line-handler process in the same processor pair as the SWAN concentrator. For information about the SWAN concentrator, refer to the WAN Subsystem Configuration and Management Manual.
PAGE 120
Configuring Direct-Connect and Satellite-Connect Lines Summary of Configuration Steps Summary of Configuration Steps Once all hardware and software requirements have been met (refer to Required Hardware and Software on page 7-2 for details), configuring and starting a single-line direct-connect or satellite-connect line-handler process involves the following steps: Step Tool Used Step 1: Find an Available WAN Line SCF interface to the WAN subsystem Step 2: Create a Profile for the Line-Handler Process
PAGE 121
Configuring Direct-Connect and Satellite-Connect Lines Step 2: Create a Profile for the Line-Handler Process Example 7-1. SCF STATUS ADAPTER Command -> status adapter $zzwan.#*, sub all WAN Manager STATUS ADAPTER for ADAPTER State........... STARTED \NODEA.$ZZWAN.#S01 Number of clips. 3 Clip 1 status : CONFIGURED Clip 2 status : CONFIGURED Clip 3 status : CONFIGURED WAN Manager STATUS SERVER for CLIP \NODEA.$ZZWAN.#S01.1 State :......... STARTED Path A..........: CONFIGURED Path B..........
PAGE 122
Configuring Direct-Connect and Satellite-Connect Lines ADD Profile Command This subsection describes how to create a profile using PEXQSSWN and PEXQSSAT. Note. Different profiles are provided for direct-connect and satellite-connect lines that are part of a multi-line path; these profiles are described in Section 13, Configuring Multi-Line Paths. ADD Profile Command To create a profile from the PEXQSSWN or PEXQSSAT profile template, use the WAN subsystem SCF ADD PROFILE command.
PAGE 123
Configuring Direct-Connect and Satellite-Connect Lines Step 3: Create the Line-Handler Process In the next example, a profile named SLHDIR is created for a single-line direct-connect line-handler process using the PEXQSSWN profile. The CLOCKSPEED_56000 modifier is set in the profile. -> ADD PROFILE $ZZWAN.#SLHDIR, FILE $SYSTEM.SYS01.
PAGE 124
Configuring Direct-Connect and Satellite-Connect Lines ADD DEVICE Command CPU cpunumber indicates the processor where this Expand line-handler process will normally execute. HP recommends that you specify the same processor as that configured for the preferred NonStop TCP/IP process used by the SWAN concentrator specified by concname. ALTCPU altcpunumber indicates the processor where the backup Expand line-handler process will normally execute.
PAGE 125
Configuring Direct-Connect and Satellite-Connect Lines Considerations initialization of the Expand line-handler process. The path will not be operational until you alter NEXTSYS to a valid value using either the WAN subsystem SCF ALTER DEVICE command or the Expand subsystem SCF ALTER PATH command. modifier_keyword is the name of an optional modifier in profile_name. modifier_keyword is added to the device record for this Expand line-handler process.
PAGE 126
Configuring Direct-Connect and Satellite-Connect Lines Step 4: Start the Line-Handler Process In the last example, a device named $DIR2 is created for a direct-connect line-handler process that uses a SWAN concentrator named S02. $DIR2 is also a member of a multi-CPU path. The SUPERPATH_ON and L4EXTPACKETS_ON modifiers are required for line-handler processes that are part of a multi-CPU path. The L4CONGCTRL_ON modifier is recommended for Expand line-handler processes that are part of a multi-CPU path.
PAGE 127
Configuring Direct-Connect and Satellite-Connect Lines Profile Modifiers Profile Modifiers This subsection lists the modifiers provided for configuring special features. It also describes default values and value ranges for the modifiers contained in the PEXQSSWN and PEXQSSAT profiles. Note. Different profiles are provided for direct-connect and satellite-connect lines that are part of a multi-line path; these profiles are described in Section 13, Configuring Multi-Line Paths.
PAGE 128
Configuring Direct-Connect and Satellite-Connect Lines PEXQSSWN and PEXQSSAT Modifiers Table 7-1.
PAGE 129
Configuring Direct-Connect and Satellite-Connect Lines PEXQSSWN and PEXQSSAT Modifiers Table 7-1. PEXQSSWN and PEXQSSAT Modifiers (page 3 of 3) Modifier Default Value Range of Values OSSPACE 32767 3072 through 32767 OSTIMEOUT 300 10 through 32767 PATHBLOCKBYTES 0 0 through 4095 PATHPACKETBYTES 1024 0 through 4095 PATHTF 0 0 through 186 PROGRAM $SYSTEM.CSSnn. C1097P00 (directconnect) $SYSTEM.CSSnn.
PAGE 130
Configuring Direct-Connect and Satellite-Connect Lines PEXQSSWN and PEXQSSAT Modifiers Expand Configuration and Management Manual—529522-002 7- 16
PAGE 131
8 Configuring Expand-Over-IP Lines The Expand-over-IP line-handler process provides connectivity to an Internet Protocol (IP) network. The Expand-over-IP line-handler process uses the services of the NonStop TCP/IP subsystem to provide Expand-over-IP connections. NonStop TCP/IPv6 supports IP version 6 (IPv6) communications. IPv6 supports a larger, 128-bit (16-byte) IP address that helps to address the growing number of machines and devices on the Internet, and it supports logical network partitioning (LNP).
PAGE 132
Required Hardware and Software Configuring Expand-Over-IP Lines Required Hardware and Software Several hardware and software components are required in addition to the Expandover-IP line-handler process to provide Expand-over-IP connectivity. Figure 8-1 shows the relationship between the Expand subsystem, the NonStop TCP/IP subsystem and the LAN adapter. The TCP/IP process in this configuration can be either a NonStop TCP/IP or NonStop TCP/IPv6 (TCP6SAM process).
PAGE 133
QIO Subsystem Configuring Expand-Over-IP Lines Figure 8-2 illustrates the required components when an ATM 3 ServerNet adapter (ATM3SA) is used to provide connectivity to an IP network. In this configuration, the TCP/IP process can only be NonStop TCP/IP; NonStop TCP/IPv6 does not support ATM communications. Figure 8-2. Expand-Over-IP Connectivity Components with ATM3SA Processor QIO Shared Memory Segment Expand-over-IP Line-Handler Process TCP/IP Process ATM Subsystem Y-Fabric X-Fabric ATM3SA VST058.
PAGE 134
Configuring Expand-Over-IP Lines NonStop TCP/IP Process NonStop TCP/IP Process The Expand-over-IP line-handler process uses the services of a NonStop TCP/IP process to provide TCP/IP connectivity. The NonStop TCP/IP process and SUBNET associated with the Expand-over-IP line-handler process must be defined and started before the Expand-over-IP line-handler process can be started. It must be configured in the same processor pair as the Expand-over-IP line-handler process.
PAGE 135
Configuring Expand-Over-IP Lines Local Area Network (LAN) Driver and Interrupt Handlers (DIHs) adapter/line may be used for all connections to the same neighbor. NonStop TCP/IPv6 (in either LNP or non-LNP mode) can be configured with a redundancy feature at the adapter level called Ethernet failover. This feature allows you to configure two network interfaces (IP addresses and their associated physical interfaces on the Ethernet adapter) as a failover pair.
PAGE 136
Configuring Expand-Over-IP Lines Topology Considerations The Gigabit Ethernet ServerNet adapter (GESA) is a single-port ServerNet adapter that provides Gigabit connectivity between Integrity NonStop NS-series systems and Ethernet LANs. A GESA can be directly installed in slots 51 through 54 of an I/O enclosure and slots 53 and 54 of a processor enclosure.
PAGE 137
Topology Considerations Configuring Expand-Over-IP Lines Figure 8-3. Expand-Over-IP Line-Handler Process Topology Node \B Node \A CPU 0 CPU 0 LH CPU 1 CPU 2 LH IP Network LH Single-Line Path LH IP Network Multi-CPU Path LH CPU 1 LH CPU 2 LH IP Network CPU 3 Node \C LH CPU 0 Multiline Path Node \D VST042.
PAGE 138
Configuring Expand-Over-IP Lines Summary of Configuration Steps Summary of Configuration Steps Once all hardware and software requirements have been met (refer to Required Hardware and Software on page 8-2 for details), configuring and starting a single-line Expand-over-IP line-handler process involves the following steps. Use Steps A or B depending on which version of TCP/IP you want to use.
PAGE 139
Step 1 (A): Select a Process and SUBNET for NonStop TCP/IP Use Configuring Expand-Over-IP Lines Step 1 (A): Select a Process and SUBNET for NonStop TCP/IP Use Note. The following instructions assume that a NonStop TCP/IP process has already been created. For information about creating NonStop TCP/IP processes, refer to the TCP/IP Configuration and Management Manual. A NonStop TCP/IP SUBNET associates a NonStop TCP/IP process with a connection to a network and an IP address.
PAGE 140
Creating an Ethernet SUBNET or ATM SUBNET Configuring Expand-Over-IP Lines Example 8-2. SCF INFO SUBNET Command 2-> INFO SUBNET $ZB01A.#* TCPIP Info SUBNET \NODEA.$ZB01A.* Name Devicename #LOOP0 #SN1 #SN2 \NODEB.$NOIOP \NODEA.$LAN01 \NODEA.$AM1 *IPADDRESS 127.0.0.1 172.16.35.15 172.16.192.
PAGE 141
Configuring Expand-Over-IP Lines Step 1 (B): Select a Process and SUBNET for NonStop TCP/IPv6 Use Step 1 (B): Select a Process and SUBNET for NonStop TCP/IPv6 Use Note. The following instructions assume that a NonStop TCP/IPv6 environment has already been started. For information about starting a NonStop TCP/IPv6 environment, refer to the TCP/IPv6 Configuration and Management Manual. Step 1 (B) is for use with NonStop TCP/IPv6 only.
PAGE 142
Configuring Expand-Over-IP Lines Select a SUBNET for NonStop TCP/IPv6 Use Example 8-3. SCF INFO SUBNET, DETAIL Command TCPIPV6 Detailed Info SUBNET \NODEA.$ZZTCP.#ZPTMF.* AF_INET: Name Devicename *IPADDRESS/DST_IPADDR TYPE *SUBNETMASK SN116 \NODEA.FEF0A 172.10.188.140 ETHERNET %HFFFFFF00 Trace Status ........ OFF Trace Filename ...... Interface MTU ....... 1500 ---Multicast Groups-----State--224.0.0.1 STARTED LNP... $ZB01A Index... 1 LOOP0 127.0.0.1 LOOP %HFF000000 Trace Status ........
PAGE 143
Select a TCP6SAM Process Configuring Expand-Over-IP Lines If you want to use the default LNP, select a SUBNET that has DEFAULT in the LNP field. Select a TCP6SAM Process The TCP/IP socket access method (TCP6SAM) is the process that provides access to the NonStop TCP/IPv6 environment. In configuring Expand-over-IP for this environment, you use the name of a TCP6SAM process for the ASSOCIATEDEV modifier in Step 4: Create the Line-Handler Process.
PAGE 144
Configuring Expand-Over-IP Lines Creating an Ethernet Subnet 1. Issue the SCF INFO SUBNET $ZZTCP.*, DETAIL command. 2. Identify all TCP6SAM processes that are listed in the LNP field of the SUBNET display and make a note of these process names. 3. Issue the SCF LISTDEV TCPIP command. 4. Use your list of TCP6SAM names to eliminate the LNP-assigned TCP6SAM processes. The remaining TCP6SAM process(es) is associated with the default LNP. This process has access only to the SUBNETs in the default partition.
PAGE 145
Step 2 (B): Identify an Available UDP Port Number for TCP/IPv6 Use Configuring Expand-Over-IP Lines You can use the SCF STATUS PROCESS command to determine which UDP port numbers are already in use for a particular SUBNET. Example 8-5 shows an example of a SCF STATUS PROCESS command for the TCP/IP process named $ZTC01: Example 8-5. SCF STATUS PROCESS Command 3-> STATUS PROCESS $ZTC01 TCPIP Status PROCESS \NODEA.$ZTC01 Status: STARTED PPID............
PAGE 146
Step 2 (B): Identify an Available UDP Port Number for TCP/IPv6 Use Configuring Expand-Over-IP Lines port number for the IP address you selected in Step 1 (B): Select a Process and SUBNET for NonStop TCP/IPv6 Use. You can use the SCF MON PROCESS command to determine which UDP port numbers are already in use for a particular SUBNET. To obtain a list of available UDP port numbers, issue the following SCF command: > STATUS MON $ZZTCP.* Example 8-6 shows a sample result of the SCF STATUS MON command.
PAGE 147
Configuring Expand-Over-IP Lines Step 3: Create a Profile for the Line-Handler Process that you do not use a well-known UDP port number in the range 0 to 1023. You must specify this UDP port number when you define the Expand-over-IP line-handler process in Step 4: Create the Line-Handler Process. Note. You must also perform this step on the destination system before you can define the local Expand-over-IP line-handler process.
PAGE 148
Configuring Expand-Over-IP Lines Example modifier_value is the value you want to assign to the modifier specified by modifier_keyword. Specifying a modifier_value assigns a new value to modifier_keyword in profile_name. Default values and ranges of values for modifiers in the PEXQSIP profile are described in Profile Modifiers on page 8-25. Example In the following example, a profile named SLHIP is created for a single-line Expandover-IP line-handler process using the PEXQSIP profile.
PAGE 149
Configuring Expand-Over-IP Lines ADD DEVICE Command $ZZWAN.#device_name specifies, via the WAN subsystem, the device name of the Expand line-handler process to add. IOPOBJECT $SYSTEM.SYSnn.LHOBJ is the name of the object file containing the executable object for code for an Expand line-handler process. This value must be $SYSTEM.SYSnn.LHOBJ. PROFILE profile_name is the name of the profile you created for this Expand line-handler process in Step 3: Create a Profile for the Line-Handler Process.
PAGE 150
Configuring Expand-Over-IP Lines ADD DEVICE Command {IPVER_IPV4 | IPVER_IPV6} specifies whether the destination and source addresses are IPv4 or IPv6. The default is IPv4. If IPVER is IPV4 (the default), then DESTIPADDR and SRCIPADDR are required. If IPVER is IPV6, then V6DESTIPADDR and V6SRCIPADDR are required. (This attribute applies to NonStop TCP/IPv6 only.
PAGE 151
Configuring Expand-Over-IP Lines Considerations V6DESTIPADDR v6destip-address if IPVER is IPv6, this is a required modifier that specifies the IP address used by the remote (destination) Expand-over-IP line-handler process. It is the IP address specified in the remote line-handler process’ V6SRCIPADDR modifier. v6dest_ipaddr must be specified by number (for example, 1611:1071:F881:1167:1611:A071:1881:B167).
PAGE 152
Configuring Expand-Over-IP Lines Step 5: Start the Line-Handler Process Examples In the following example, a device named $IPLIN1 is created for a single-line Expandover-IP line-handler process. The PATHPACKETBYTES modifiers are recommended for Expand-over-IP lines. (It is not necessary to specify the L4EXTPACKETS_ON and L4CONGCTRL_ON modifiers since they are always ON, per Meena.) Per Dave, “PATHBLOCKBYTES is not so easy.
PAGE 153
Configuring Expand-Over-IP Lines Step 6: Start the Line This command creates the specified Expand line-handler process and allocates a logical device (LDEV) number. Step 6: Start the Line To start an Expand-over-IP line, use the Expand subsystem SCF START LINE command. The command syntax is as follows: START LINE $device_name device_name is the device name of the Expand-over-IP line-handler process. The successful completion of this command leaves the line in the STARTED state.
PAGE 154
Configuring Expand-Over-IP Lines Add a Configured Tunnel for an Expand Line Example 8-8 shows how to add an Expand line from \NodeB to \NodeC. Example 8-8. Add an Expand Line to \NodeC allow all errors abort line $giplco1 stop device $zzwan.#giplco1 delete device $zzwan.#giplco1 == Add profile of IP line ADD PROFILE $zzwan.#afkslhip, file $data00.t9057afk.sippfr == Add Expand line handler. ADD DEVICE $ZZWAN.#giplco1, CPU 2, ALTCPU 3, PROFILE afkslhip,& IOPOBJECT $data00.t9057afk.
PAGE 155
Configuring Expand-Over-IP Lines Profile Modifiers Example 8-10 shows how to add an Expand line from \NodeC to \NodeB. Example 8-10. Add an Expand Line to \NodeB allow all errors abort line $giplba1 stop device $zzwan.#giplba1 delete device $zzwan.#giplba1 == Add profile of IP line ADD PROFILE $zzwan.#afkslhip, file $data00.t9057afk.sippfr == Add Expand line handler. ADD DEVICE $ZZWAN.#giplba1, CPU 2, ALTCPU 3, PROFILE afkslhip,& IOPOBJECT $data00.t9057afk.
PAGE 156
Configuring Expand-Over-IP Lines Modifiers for Special Features L4CONGCTRL is a path parameter and the path profile sets L4CONGCTRL_OFF because it is shared by all line types. Therefore, multi-line IP paths default to L4CONGCTRL_OFF and must specify L4CONGCTRL_ON. The L4CONGCTRL_ON modifier is also recommended for Expand line-handler processes that are part of a multi-CPU path.
PAGE 157
PEXQSIP Modifiers Configuring Expand-Over-IP Lines Table 8-1. PEXQSIP Modifiers for Expand-over-IP Lines (page 1 of 2) Modifier Default Value Range of Values AFTERMAXRETRIES_DOWN AFTERMAXRETRIES_PASSIVE 3 ASSOCIATEDEV1 None Any 8-character string COMPRESS_OFF COMPRESS_ON 3 CONNECTTYPE_ACTIVEANDPASSIVE 3 CONNECTTYPE_PASSIVE DESTIPADDR 2 0.0.0.
PAGE 158
PEXQSIP Modifiers Configuring Expand-Over-IP Lines Table 8-1. PEXQSIP Modifiers for Expand-over-IP Lines (page 2 of 2) Modifier Default Value Range of Values RETRYPROBE 19 1 through 255 RXWINDOW 7 2 through 15 SPEED 0 0 through 224000 SPEEDK NOT_SET 0 through 4,000,000,000 SRCIPADDR2 0.0.0.
PAGE 159
9 Configuring Expand-Over-ATM Lines The Expand-over-ATM line-handler process provides connectivity to an Asynchronous Transfer Mode (ATM) network. In addition, the Expand-over-ATM line-handler process can use the services of the ServerNet LAN systems access (SLSA) subsystem to provide Expand-over-ATM connections. The ATM subsystem provides a PVC and SVC connection, whereas the SLSA subsystem provides an ATMSAP with lifname connection that enables you to manage PVC connections under SLSA.
PAGE 160
Required Hardware and Software Configuring Expand-Over-ATM Lines Required Hardware and Software Several hardware and software components are required in addition to the Expandover-ATM line-handler process to provide Expand-over-ATM connectivity. These components are illustrated in Figure 9-1 and are explained in the following subsections. Figure 9-1.
PAGE 161
Configuring Expand-Over-ATM Lines ATM Subsystem ATM Subsystem ATM is a cell-switching and multiplexing technology that combines the benefits of circuit switching (constant transmission delay and guaranteed capacity) with those of packet switching (flexibility and intermittent traffic). The ATM subsystem is the HP implementation of the ATM technology. The ATM subsystem supports the ATM UserNetwork Interface (UNI) Specification Version 3.0 over a 155 Mbps SONET STS-3c connection.
PAGE 162
Topology Considerations Configuring Expand-Over-ATM Lines Topology Considerations In a single-line configuration, you configure one Expand-over-ATM line-handler process for each path to an adjacent node. In a multi-CPU path configuration, you configure multiple Expand-over-ATM line-handler processes, usually in separate processors, for each path to an adjacent node. In a multi-line path configuration, you configure a path that consists of multiple lines between adjacent nodes.
PAGE 163
Summary of Configuration Steps Configuring Expand-Over-ATM Lines Summary of Configuration Steps Once all hardware and software requirements have been met (refer to Required Hardware and Software on page 9-2 for details), configuring and starting a single-line Expand-over-ATM line-handler process involves the following steps: Step Tool Used Step 1: Identify the ATM Connection SCF interface to the ATM subsystem Step 2: Create a Profile for the Line-Handler Process SCF interface to the WAN subsystem St
PAGE 164
Configuring an Expand Line-Handler Process That Uses a PVC Configuring Expand-Over-ATM Lines Configuring an Expand Line-Handler Process That Uses a PVC If your Expand-over-ATM line-handler process will use a PVC connection, you must identify the PVC you plan to use. The SCF INFO PVC command displays the configured name of a PVC. Example 9-1 shows an example of an SCF INFO PVC command for an ATM line named $AM1. Example 9-1. SCF INFO PVC Command 1-> INFO PVC $AM1.#IP.* ATM Info PVC Name $AM1.#IP.
PAGE 165
Configuring an Expand Line-Handler Process That Uses an SVC Configuring Expand-Over-ATM Lines your local and remote Expand-over-ATM line-handler processes. Specifying a selector byte when configuring an Expand-over-ATM line-handler process is described in Step 3: Create the Line-Handler Process.
PAGE 166
Configuring an Expand Line-Handler Process That Uses an SVC Configuring Expand-Over-ATM Lines Specifying an ATM address when configuring an Expand-over-ATM line-handler process in described in Step 3: Create the Line-Handler Process.
PAGE 167
Configuring an Expand Line-Handler Process That Uses ATMSAP Configuring Expand-Over-ATM Lines Configuring an Expand Line-Handler Process That Uses ATMSAP The SLSA ATMSAP connection offers an ATM Native Mode network interconnect support similar to that offered by the PVC object within the ATM subsystem. Expand issues native mode frames directly to the ATM product via a LIF associated with an ATMSAP object. Figure 9-3 illustrates ATMSAP use by Expand. Figure 9-3.
PAGE 168
Step 2: Create a Profile for the Line-Handler Process Configuring Expand-Over-ATM Lines Verifying the Line-Handler Process Example 9-4 shows an example of an Expand subsystem SCF INFO LINE command with the DETAIL option of an Expand-over-ATM line-handler process named $ATM2BAT and a CallType of ATMSAP. Example 9-4. Expand Subsystem SCF INFO LINE, DETAIL Command for ATMSAP -> INFO LINE $ATM2BAT, DETAIL EXPAND Detailed Info L2Protocol Framesize.... *LinePriority... *DownIfBadQuality *Txwindow...
PAGE 169
Configuring Expand-Over-ATM Lines ADD Profile Command ADD Profile Command To create a profile from the PEXQSATM profile template, use the WAN subsystem SCF ADD PROFILE command. The command syntax is as follows: ADD PROFILE $ZZWAN.#profile_name , FILE $SYSTEM.SYSnn.profile_filename [, modifier_keyword [ modifier_value ] ] ... $ZZWAN.profile_name specifies, via the WAN subsystem, a user-defined name of up to eight alphanumeric characters that will be used to identify the new profile.
PAGE 170
Configuring Expand-Over-ATM Lines Step 3: Create the Line-Handler Process Step 3: Create the Line-Handler Process You create a single-line Expand-over-ATM line-handler process by adding it as a device to the WAN subsystem. Note. This section explains how to configure single-line Expand-over-ATM line-handler processes only. Creating an Expand-over-ATM line that is part of a multi-line path is explained in Section 13, Configuring Multi-Line Paths.
PAGE 171
Configuring Expand-Over-ATM Lines ADD DEVICE Command Syntax for SVC Connections Use the following command syntax if the Expand-over-ATM line-handler process will use an SVC connection: ADD , , , , , , , , , , , , [, DEVICE $ZZWAN.#device_name IOPOBJECT $SYSTEM.SYSTEM.
PAGE 172
Configuring Expand-Over-ATM Lines ADD DEVICE Command CPU cpunumber indicates the processor where this Expand line-handler process will normally execute. ALTCPU altcpunumber indicates the processor where the backup Expand line-handler process will normally execute. TYPE (63,0) is the device type and subtype for this Expand line-handler process. The device type is always 63 for Expand line-handler processes. The subtype is 0 for singleline Expand-over-ATM line-handler processes.
PAGE 173
Configuring Expand-Over-ATM Lines ADD DEVICE Command PVCNAME pvc-name is the name of the permanent virtual circuit (PVC) that will be used. This is the PVC name you identified in Step 1: Identify the ATM Connection on page 9-5. For example, PVC01. This modifier is only applicable to Expand-over-ATM line-handler processes that use PVC connections. ATMSEL selector-byte is a hexadecimal selector byte for the ATM line used by this Expand-over-ATM line-handler process.
PAGE 174
Configuring Expand-Over-ATM Lines Considerations modifier_value is the value you want to assign to the optional modifier specified by modifier_keyword. modifier_value assigns a value to modifier_keyword in the device record for this Expand line-handler process. Default values and ranges of values for modifiers in the PEXQSATM profile are described in Profile Modifiers on page 9-17. Considerations • • Not all modifiers have associated values (for example, L4EXTPACKETS_ON).
PAGE 175
Configuring Expand-Over-ATM Lines Step 4: Start the Line-Handler Process In the last example, a device named ATMLIN3 is created for a single-line Expand-overATM line-handler process that uses an ATMSAP connection through the SLSA subsystem. -> ADD DEVICE $ZZWAN.#ATMLIN3, PROFILE SLHATM, IOPOBJECT & $SYSTEM.SYSTEM.
PAGE 176
Configuring Expand-Over-ATM Lines Recommended Modifiers Recommended Modifiers Recommended modifiers are modifiers that should be used to obtain optimum performance and efficiency. L4CONGCTRL_ON Default: Units: Range: ON Not applicable ON or OFF This modifier enables the congestion control mechanism. Because data transfer with the UDP is not guaranteed, the Expand End-to-End protocol is used to achieve reliable communications for Expand-over-ATM connections.
PAGE 177
Modifiers for Special Features Configuring Expand-Over-ATM Lines Modifiers for Special Features In addition to the L4CONGCTRL_ON and PATHPACKETBYTES modifiers, the SUPERPATH_ON modifier is provided in the PEXQSATM profile to enable you to configure the Expand multi-CPU feature. For configuration considerations for all special features, refer to Section 17, Subsystem Description. For information about the advantages and disadvantages of each feature, refer to Section 3, Planning a Network Design.
PAGE 178
PEXQSATM Modifiers Configuring Expand-Over-ATM Lines Table 9-1.
PAGE 179
PEXQSATM Modifiers Configuring Expand-Over-ATM Lines Table 9-1. PEXQSATM Modifiers for Expand-over-ATM Lines (page 3 of 3) Modifier Default Value Range of Values TIMERRECONNECT 30 30 through 32767 TXWINDOW 7 2 through 25 1. This is a required modifier. 2. This modifier is required for Expand-over-ATM line-handler processes that use SVC connections. 3. This modifier is required for Expand-over-ATM line-handler processes that use PVC connections. 4.
PAGE 180
Configuring Expand-Over-ATM Lines Expand Configuration and Management Manual—529522-002 9- 22 PEXQSATM Modifiers
PAGE 181
10 Configuring Expand-Over-X.25 Lines X.25 is a standard for private and public networks that use packet-switching technology. Expand-over-X.25 connections are provided by the HP X.25 Access Method (X25AM) product. The Expand-over-X.25 line-handler process uses the NETNAM protocol to access the network access method (NAM) interface provided by an X25AM line-handler process. An Expand-over-X.
PAGE 182
Required Hardware and Software Configuring Expand-Over-X.25 Lines Figure 10-1. Expand-Over-X.25 Line-Handler Process Components Processor Expand-over-X.
PAGE 183
Configuring Expand-Over-X.25 Lines X25AM Line-Handler Process X25AM Line-Handler Process The Expand-over-X.25 line-handler process uses the services of an X25AM line-handler process to provide access to X.25 packet-switched data networks (PSDNs). Each X25AM line-handler process controls a single data communications line and supports both permanent virtual circuits (PVCs) and switched virtual circuits (SVCs). The X25AM line-handler process associated with the Expand-over-X.
PAGE 184
Configuring Expand-Over-X.25 Lines Local Area Network (LAN) Driver and Interrupt Handlers (DIHs) Local Area Network (LAN) Driver and Interrupt Handlers (DIHs) NonStop TCP/IP processes can interface to the network through the ServerNet LAN Systems Access (SLSA) subsystem. The SLSA subsystem provides QIO-based driver and interrupt handlers (DIHs) that allow NonStop TCP/IP processes to connect to a LAN adapter. The SLSA subsystem is preconfigured and started during the system load sequence.
PAGE 185
Configuring Expand-Over-X.25 Lines ServerNet Wide Area Network (SWAN) Concentrator the IOAM enclosure, and it cannot be installed in an Integrity NonStop NS-series enclosure. The data transfer rate is limited to 10 Mbps when the G4SA is connected to a SWAN concentrator, but a data transfer rate of 10/100 Mbps is possible when the G4SA is connected to a SWAN 2 concentrator. For information about the G4SA, refer to the LAN Configuration and Management Manual.
PAGE 186
Topology Considerations Configuring Expand-Over-X.25 Lines Figure 10-2. Expand-Over-X.25 Line-Handler Process Topology Node \A CPU 0 LH CPU 1 CPU 2 Single-Line Path LH LH Node \C CPU 0 Multi-CPU Path PSDN LH LH LH CPU 0 LH CPU 1 CPU 2 Multiline Path LH Node \B Key Configured Path CDT 043.
PAGE 187
Summary of Configuration Steps Configuring Expand-Over-X.25 Lines Summary of Configuration Steps Once all hardware and software requirements have been met (refer to Required Hardware and Software on page 10-1 for details), configuring and starting a single-line Expand-over-X.25 line-handler process involves the following steps.
PAGE 188
Configuring Expand-Over-X.25 Lines Considerations For more information, refer to the X25AM Configuration and Management Manual. Considerations The following configuration considerations apply to SU objects: • • • The PROTOCOL attribute identifies the protocol that will be used by the subdevice. Subdevices used by Expand-over-X.25 line-handler processes must use the NETNAM protocol (specified by the argument NAM). The DEVTYPE attribute specifies the subdevice type.
PAGE 189
Configuring Expand-Over-X.25 Lines ADD Profile Command ADD Profile Command To create a profile from the PEXQSNAM profile template, use the WAN subsystem SCF ADD PROFILE command. The command syntax is as follows: ADD PROFILE $ZZWAN.#profile_name , FILE $SYSTEM.SYSnn.profile_filename [, modifier_keyword [ modifier_value ] ] ... $ZZWAN.profile_name specifies, via the WAN subsystem, a user-defined name of up to 8 alphanumeric characters that will be used to identify the new profile.
PAGE 190
Configuring Expand-Over-X.25 Lines Step 4: Create the Expand-Over-X.25 Line-Handler Process Step 4: Create the Expand-Over-X.25 Line-Handler Process You create a single-line Expand-over-X.25 line-handler process by adding it as a device to the WAN subsystem. Note. This section explains how to configure single-line Expand-over-X.25 line-handler processes only. Creating Expand-over-X.25 lines that are part of a multi-line path is explained in Section 13, Configuring Multi-Line Paths.
PAGE 191
Configuring Expand-Over-X.25 Lines ADD DEVICE Command ALTCPU altcpunumber indicates the processor where the backup Expand line-handler process will normally execute. TYPE (63,0) is the device type and subtype for this Expand line-handler process. The device type is always 63 for Expand line-handler processes. The subtype is 0 for singleline Expand-over-X.25 line-handler processes. RSIZE rsize specifies the time factor of the line for the Expand routing algorithm.
PAGE 192
Configuring Expand-Over-X.25 Lines Considerations Default values and ranges of values for modifiers in the PEXQSNAM profile are described in Profile Modifiers on page 10-13. Considerations • • Not all modifiers have associated values (for example, L4EXTPACKETS_ON). The modifier_keyword and modifier_value parameters do not add the specified modifier, or a modifier and its associated value, to the profile used by the device.
PAGE 193
Configuring Expand-Over-X.25 Lines Step 6: Start the Expand-Over-X.25 Line Step 6: Start the Expand-Over-X.25 Line To start an Expand-over-X.25 line, use the Expand subsystem SCF START LINE command. The command syntax is as follows: START LINE $device_name device_name is the device name of the Expand-over-X.25 line-handler process. The successful completion of this command leaves the line in the STARTED state. Profile Modifiers This subsection lists the recommended modifiers for Expand-over-X.
PAGE 194
Configuring Expand-Over-X.
PAGE 195
PEXQSNAM Modifiers Configuring Expand-Over-X.25 Lines Table 10-1. PEXQSNAM Modifiers for Expand-over-X.
PAGE 196
PEXQSNAM Modifiers Configuring Expand-Over-X.25 Lines Table 10-1. PEXQSNAM Modifiers for Expand-over-X.25 Lines (page 2 of 2) Modifier Default Value Range of Values TIMERINACTIVITY 900 0 through 32767 TIMERPROBE 300 1 through 32767 TIMERRECONNECT 30 0 through 32767 TXWINDOW 4 2 through 7 SUPERPATH_ON 1. This is a required modifier. It has no default value. 2. This is a required modifier. The default value is invalid and must be changed.
PAGE 197
11 Configuring Expand-Over-SNA Lines Systems Network Architecture (SNA) was developed by IBM for connecting IBM systems and networks. Expand-over-SNA connections are provided with the HP SNAX/Advanced Peer Networking (SNAX/APN) product. The Expand-over-SNA line-handler process uses the NETNAM protocol to access the network access method (NAM) interface provided by a SNAX/APN line-handler process.
PAGE 198
Required Hardware and Software Configuring Expand-Over-SNA Lines Figure 11-1.
PAGE 199
Configuring Expand-Over-SNA Lines SNAX/APN Line-Handler Process SNAX/APN Line-Handler Process The Expand-over-SNA line-handler process uses the services of a SNAX/APN line-handler process to provide access to an IBM SNA network. The SNA network can be a traditional network of host mainframes and front end processors, an advanced peer-to-peer network of AS400 systems or other workstations, or a mix of these types of networks.
PAGE 200
Configuring Expand-Over-SNA Lines NonStop TCP/IP Process NonStop TCP/IP Process The NonStop TCP/IP subsystem provides TCP/IP data communications connectivity. NonStop TCP/IP processes are used by the following LAN adapters: Ethernet 4 ServerNet adapters (E4SAs), Fast Ethernet ServerNet adapters (FESAs), Gigabit Ethernet ServerNet adapters (GESAs), Gigabit Ethernet 4-port ServerNet adapters (G4SAs), and SWAN concentrators.
PAGE 201
Configuring Expand-Over-SNA Lines Gigabit Ethernet ServerNet Adapter (GESA) Gigabit Ethernet ServerNet Adapter (GESA) The Gigabit Ethernet ServerNet adapter (GESA) is a single-port ServerNet adapter that provides Gigabit connectivity between Integrity NonStop NS-series systems and Ethernet LANs. A GESA can be directly installed in slots 51 through 54 of an I/O enclosure and slots 53 and 54 of a processor enclosure.
PAGE 202
Topology Considerations Configuring Expand-Over-SNA Lines Figure 11-2. Expand-Over-SNA Line-Handler Process Topology Node \A CPU 0 LH CPU 1 CPU 2 Single-Line Path LH LH Node \C CPU 0 Multi-CPU Path SNA Network LH LH LH CPU 0 LH CPU 1 CPU 2 Multiline Path LH Node \B Key Configured Path CDT 044.
PAGE 203
Configuring Expand-Over-SNA Lines Summary of Configuration Steps Summary of Configuration Steps Once all hardware and software requirements have been met (refer to Required Hardware and Software on page 11-1 for details), configuring and starting a single-line Expand-over-SNA line-handler process involves the following steps: Step Tool Used Step 1: Add the SNAX/APN Line SCF interface to the SNAX/APN subsystem Step 2: Add the LUs for the SNAX/APN Line SCF interface to the SNAX/APN subsystem Step 3: S
PAGE 204
Configuring Expand-Over-SNA Lines Step 1: Add the SNAX/APN Line Step 1: Add the SNAX/APN Line Note. The following instructions assume that a SNAX/APN line-handler process has already been created. To create such a process, you must add it as a device to the WAN subsystem. For information about creating SNAX/APN line-handler processes, refer to the WAN Subsystem Configuration and Management Manual.
PAGE 205
Configuring Expand-Over-SNA Lines Considerations For details about configuring LUs and PUs, and SCF ADD LU and ADD PU command attributes, refer to the SNAX/XF and SNAX/APN Configuration and Management Manual. Considerations The following configuration considerations apply to local LU object attributes: • • • The PROTOCOL attribute must be set to NAM. The local SNANAME on one system must match the remote LU SNANAME on the other system.
PAGE 206
Example Configuring Expand-Over-SNA Lines Figure 11-3. SNAX/APN Line Configuration Example System \A Expand Line-Handler Process SNAX/APN Line-Handler Process ($SNAPA) SWAN SNA Network SCF Commands for System \A ADD LINE $SNAPA , RECSIZE 524 , MAXPUS 1 , MAXLUS 30 , STATION PRIMARY , MAXLOCALLUS 10 , POLLINT 0.01 ADD LU $SNAPA.#LLUA , TYPE (14,21) , SNANAME LUA , PROTOCOL NAM , DLUNAME #RLUA , RSPTYPE ER ADD PU $SNAPA.#RPUA , TYPE (13,21) , ADDRESS %HC1 , RECSIZE 521 , MAXLUS 30 ADD LU $SNAPA.
PAGE 207
Configuring Expand-Over-SNA Lines Step 3: Start the SNAX/APN Line Step 3: Start the SNAX/APN Line Before you can start the Expand-over-SNA line, the SNAX/APN line (and its associated PUs and LUs) must be started. To start a SNAX/APN line, use the SNAX/APN subsystem SCF START LINE command.
PAGE 208
Configuring Expand-Over-SNA Lines Example modifier_keyword is the name of a modifier in profile_name. Modifier names in the PEXQSNAM profile are listed in Profile Modifiers on page 11-16. modifier_value is the value you want to assign to the modifier specified by modifier_keyword. Specifying a modifier_value assigns a new value to modifier_keyword in profile_name. Default values and ranges of values for modifiers in the PEXQSNAM profile are described in Profile Modifiers on page 11-16.
PAGE 209
Configuring Expand-Over-SNA Lines ADD DEVICE Command $ZZWAN.#device_name specifies, via the WAN subsystem, the device name of the Expand line-handler process to add. IOPOBJECT $SYSTEM.SYSnn.LHOBJ is the name of the object file containing the executable object for code for an Expand line-handler process. This value must be $SYSTEM.SYSnn.LHOBJ.
PAGE 210
Configuring Expand-Over-SNA Lines Considerations NEXTSYS sys_number is a required modifier that specifies the number (from 0 through 254) of the system connected to the other end of the line. If you do not specify NEXTSYS, this modifier defaults to an invalid value (255) and an operator message occurs during the initialization of the Expand-over-SNA line-handler process.
PAGE 211
Configuring Expand-Over-SNA Lines Step 6: Start the Expand-Over-SNA Line-Handler Process In the next example, the same device is created as part of a multi-CPU path. The SUPERPATH_ON and L4EXTPACKETS_ON modifiers are required for line-handler processes that are part of a multi-CPU path. The L4CONGCTRL_ON modifier is recommended for Expand line-handler processes that are part of a multi-CPU path. -> ADD DEVICE $ZZWAN.#EXPS14, PROFILE SLHSNA, IOPOBJECT & $SYSTEM.SYSTEM.
PAGE 212
Configuring Expand-Over-SNA Lines Profile Modifiers Profile Modifiers This subsection lists the recommended modifiers for Expand-over-SNA line-handler processes and describes the modifiers provided for configuring special features. It also describes default values and value ranges for all the modifiers contained in the PEXQSNAM profile. Note. A different profile is provided for Expand-over-SNA lines that are part of a multi-line path; this profile is described in Section 13, Configuring Multi-Line Paths.
PAGE 213
PEXQSNAM Modifiers Configuring Expand-Over-SNA Lines The PATHBLOCKBYTES, PATHPACKETBYTES, L4CONGCTRL_ON, and SUPERPATH_ON modifiers are described in detail in Section 16, Expand Modifiers. PEXQSNAM Modifiers The disk file $SYSTEM.SYSnn.PEXQSNAM defines modifiers for Expand-over-SNA line-handler processes. Table 11-1 lists the default value and range of values for each modifier in this profile, if applicable.
PAGE 214
PEXQSNAM Modifiers Configuring Expand-Over-SNA Lines Table 11-1.
PAGE 215
12 Configuring Expand-Over-ServerNet Lines The Expand-over-ServerNet line-handler process provides connectivity to a ServerNet Cluster, which uses this process to provide a high-speed interconnect between systems over a limited geographic range. The Expand-over-ServerNet line-handler process uses the NETNAM protocol to access the network access method (NAM) interface of the ServerNet cluster monitor process, $ZZSCL.
PAGE 216
Configuring Expand-Over-ServerNet Lines Expand Manager Process ($ZEXP) Figure 12-1 depicts a local node consisting of three processors: Processor 1 is running an application, Procesor 2 is running the ServerNet cluster monitor process ($ZZSCL), and Processor 3 is running an Expand-over-ServerNet line handler. The application makes a communications request to the message system.
PAGE 217
Network Access Method (NAM) Configuring Expand-Over-ServerNet Lines MSGMON is a persistent process. Once it is started, it terminates only in the event of an internal failure or a termination message from the persistence monitor, $ZPM. MSGMON is not a process pair. Note. MSGMON is compatible only with G06.09 and later RVUs of the NonStop Kernel operating system. Network Access Method (NAM) NAM is a pair of message system dialects that allow use of another process, such as X.
PAGE 218
Configuring Expand-Over-ServerNet Lines ServerNet Cluster Monitor Process ($ZZSCL) ServerNet Cluster Monitor Process ($ZZSCL) The ServerNet cluster monitor process, $ZZSCL, monitors and responds to events relevant to ServerNet cluster operations and is responsible for discovering and managing the cluster. The ServerNet cluster monitor process is not limited to a particular processor pair; it can migrate within a user-specified set.
PAGE 219
Topology Considerations Configuring Expand-Over-ServerNet Lines Topology Considerations A ServerNet cluster must be configured as a logical fully connected mesh—each server must have one Expand-over-ServerNet line-handler process for each other node in the ServerNet cluster. A ServerNet cluster can consist of up to 24 nodes. Figure 12-2 is an example of Expand-over-ServerNet lines in a four-node ServerNet cluster configuration. Figure 12-2.
PAGE 220
Summary of Configuration Steps Configuring Expand-Over-ServerNet Lines Summary of Configuration Steps Once all hardware and software requirements have been met (refer to Required Hardware and Software on page 12-1 for details), configuring Expand-over-ServerNet connections involves the following steps: Step Tool Used Step 1: Create a Profile for the Expand-OverServerNet Line-Handler Process SCF interface to the WAN subsystem Step 2: Create a Device for the Expand-OverServerNet Line-Handler Process SC
PAGE 221
Configuring Expand-Over-ServerNet Lines ADD Profile Command ADD Profile Command To create a profile from the PEXPSSN profile template, use the WAN subsystem SCF ADD PROFILE command. The command syntax is as follows: ADD PROFILE $ZZWAN.#profile_name , FILE $SYSTEM.SYSnn.profile_filename [, modifier_keyword [ modifier_value ] ] ... $ZZWAN.profile_name specifies, via the WAN subsystem, a user-defined name of up to eight alphanumeric characters that will be used to identify the new profile.
PAGE 222
Configuring Expand-Over-ServerNet Lines Step 2: Create a Device for the Expand-OverServerNet Line-Handler Process Step 2: Create a Device for the Expand-OverServerNet Line-Handler Process You create an Expand-over-ServerNet line-handler process by adding it as a device to the WAN subsystem. For each system you add to the ServerNet cluster, you must configure line-handler processes for all of the other systems in the cluster.
PAGE 223
Configuring Expand-Over-ServerNet Lines ADD DEVICE Command ALTCPU altcpunumber indicates the processor where the backup Expand-over-ServerNet line-handler process will normally execute. TYPE (63,4) is the device type and subtype for Expand-over-ServerNet line-handler processes. The device type is always 63 for Expand line-handler processes. The subtype is always 4 for Expand-over-ServerNet line-handler processes. RSIZE rsize specifies the time factor of the line for the Expand routing algorithm.
PAGE 224
Configuring Expand-Over-ServerNet Lines Considerations Considerations • • • Not all modifiers have associated values (for example, L4EXTPACKETS_ON). The modifier_keyword and modifier_value parameters do not add the specified modifier, or a modifier and its associated value, to the profile used by the device. Use the ADD PROFILE command to add a modifier, or a modifier and its associated value, to a profile. Here are recommended configuration guidelines for configuring Expand-overServerNet line handlers.
PAGE 225
Configuring Expand-Over-ServerNet Lines Step 3: Start the Expand-Over-ServerNet LineHandler Processes Step 3: Start the Expand-Over-ServerNet LineHandler Processes To start an Expand-over-ServerNet line-handler process, use the WAN subsystem SCF START DEVICE command. You must start each Expand-over-ServerNet linehandler process that you created in Step 2: Create a Device for the Expand-OverServerNet Line-Handler Process. The command syntax is as follows: START DEVICE $ZZWAN.#device_name $ZZWAN.
PAGE 226
Profile Modifiers Configuring Expand-Over-ServerNet Lines For the next steps, such as installing a new cluster, migration, or adding a node, see the ServerNet Cluster Manual or the ServerNet Cluster 6780 Planning and Installation Guide . Profile Modifiers This subsection lists the modifiers provided for configuring special features. It also describes default values and value ranges for all the modifiers contained in the PEXPSSN profile.
PAGE 227
PEXPSSN Modifiers Configuring Expand-Over-ServerNet Lines Table 12-2.
PAGE 228
Configuring Expand-Over-ServerNet Lines Expand Configuration and Management Manual—529522-002 12 -14 PEXPSSN Modifiers
PAGE 229
13 Configuring Multi-Line Paths The Expand multi-line path feature enables you to configure as many as eight lines between the two adjacent nodes. The Expand subsystem can simultaneously transmit data over all the lines in a multi-line path, thus increasing overall bandwidth, and will automatically retransmit data over remaining lines if one or more lines fail. A multi-line path can be part of a multi-CPU path. This section explains how to configure the Expand multi-line path feature. Note.
PAGE 230
Configuration Considerations Configuring Multi-Line Paths Configuration Considerations Consider the following when configuring a multi-line path: • • • • • You can configure a maximum of eight lines in a multi-line path. The lines in a multi-line path can be all the same type (for example, all dedicated), or they can be any combination of dedicated lines, X.25 connections, and SNAX connections. You cannot mix satellite-connect, Expand-over-ATM, and Expandover-IP lines with other line types.
PAGE 231
Configuring Multi-Line Paths Step 1: Create a Profile for the Path-Logical Device Note. The SCF command syntax shown in this section is the syntax used to configure multiline paths; it is not meant to show the complete syntax of the SCF commands described.
PAGE 232
Configuring Multi-Line Paths Step 2: Create a Profile for Each Line Type modifier_value is the value you want to assign to the modifier specified by modifier_keyword. Specifying a modifier_value assigns a new value to modifier_keyword in profile_name. Default values and ranges of values for modifiers in the PEXPPATH profile are described in PEXPPATH Modifiers on page 13-15. Step 2: Create a Profile for Each Line Type You must create a profile for each type of line that will be in the multi-line path.
PAGE 233
Step 3: Create a Path-Logical Device Configuring Multi-Line Paths Table 13-1. Profiles for Line-Logical Devices Disk Filename Type of Line-Logical Device PEXQMSWN Direct-connect PEXQMSAT Satellite-connect PEXQMNAM Expand-over-NAM PEXQMIP Expand-over-IP PEXQMATM Expand-over-ATM modifier_keyword is the name of a modifier in profile_name. Modifier names in the line-logical device profiles are listed in Line-Logical Device Modifiers on page 13-16.
PAGE 234
Configuring Multi-Line Paths ADD DEVICE Command IOPOBJECT $SYSTEM.SYSnn.LHOBJ is the name of the object file containing the executable object code for an Expand line-handler process. This value must be $SYSTEM.SYSnn.LHOBJ. PROFILE profile_name is the name of the profile you created for the path in Step 1: Create a Profile for the Path-Logical Device. CPU cpunumber is the processor number where the path-logical device will normally execute.
PAGE 235
Configuring Multi-Line Paths Considerations modifier_value is the value you want to assign to the optional modifier specified by modifier_keyword. modifier_value assigns a value to modifier_keyword in the device record for this path-logical device. Default values and ranges of values for modifiers in PEXPPATH are described in PEXPPATH Modifiers on page 13-15. Considerations • • Not all modifiers have associated values (for example, L4EXTPACKETS_ON).
PAGE 236
ADD DEVICE Command Configuring Multi-Line Paths PROFILE profile_name is the name of the profile you created for this type of line in Step 2: Create a Profile for Each Line Type. CPU cpunumber is the processor number where the line-logical device will normally execute. ALTCPU altcpunumber is the processor number where the backup line-logical device will normally execute. TYPE devsubtype is the device subtype for this line-logical device. The device subtypes for linelogical devices are listed in Table 13-2.
PAGE 237
Configuring Multi-Line Paths ADD DEVICE Command Required Modifiers for Direct-Connect and Satellite-Connect Lines ADAPTER concname is the ServerNet wide area network (SWAN) concentrator to be used by this line. Selecting a SWAN concentrator is explained in Step 1: Find an Available WAN Line on page 7-6. For information about adding SWAN concentrators, refer to the WAN Subsystem Configuration and Management Manual.
PAGE 238
Configuring Multi-Line Paths ADD DEVICE Command Required Modifiers for Expand-Over-IP Lines SRCIPADDR src_ipaddr is a required modifier that specifies the IP address associated with the NonStop TCP/IP process used by this Expand-over-IP line-handler process. Determining IP addresses is described in Step 1 (A): Select a Process and SUBNET for NonStop TCP/IP Use on page 8-9. The address must be specified by number (for example, 130.252.12.3). It is not validated and need not be accessible. The default is 0.
PAGE 239
Configuring Multi-Line Paths ADD DEVICE Command V6SRCIPADDR v6src_ipaddr is a required modifier that specifies the IP address associated with the NonStop TCP/IPv6 process used by this Expand-over-IP line-handler process. Determining IP addresses is described in Step 1 (B): Select a Process and SUBNET for NonStop TCP/IPv6 Use on page 8-11. The address must be specified by number (for example, 31CA:B145:5489:1034:1784:B245:4029:1257). It is not validated and need not be accessible.
PAGE 240
Configuring Multi-Line Paths Considerations ATMSEL selector_byte is a hexadecimal selector byte for the ATM line used by this Expand-over-ATM line-handler process. Obtaining selector bytes is described in Obtaining Selector Bytes for the Local and Remote ATM Lines on page 9-6. This modifier is only applicable to Expand-over-ATM line-handler processes that use SVC connections.
PAGE 241
Configuring Multi-Line Paths Step 5: Start the Path-Logical Device Step 5: Start the Path-Logical Device To start the path-logical device, use the WAN subsystem SCF START DEVICE command. When you use this command on the path-logical device, the line-logical devices associated with the path are also started. The command syntax is as follows: START DEVICE $ZZWAN.#device_name $ZZWAN.#device_name specifies, via the WAN subsystem, the path-logical device name or a line-logical device name.
PAGE 242
Configuration Example Configuring Multi-Line Paths Configuration Example The following example shows a multi-line path with one direct-connect line and one Expand-over-SNA line. Figure 13-2 illustrates the configuration of this example. Figure 13-2. Multi-Line Configuration Example MULTI $LINE1 SWAN003A $PATH $SNA1 $LINE2 MULTI SWANxxxxx ASSOCIATEDEV CDT 046.CDD Note.
PAGE 243
Configuring Multi-Line Paths • Path-Logical Device Modifiers The following SCF ADD DEVICE command creates a line-logical device named $LINE1 for the direct-connect line. Note that the MLHDIR profile created above is used. -> ADD DEVICE $ZZWAN.#LINE1, PROFILE MLHDIR, IOPOBJECT & $SYSTEM.SYSTEM.
PAGE 244
Line-Logical Device Modifiers Configuring Multi-Line Paths Table 13-3.
PAGE 245
PEXQMSWN and PEXQMSAT Modifiers Configuring Multi-Line Paths L3WINDOW n Default: Units: Range: 2 Packets 1 through 15 (L3MOD128), 1 through 7 (L3MOD8) This modifier specifies the number of packets that can be outstanding without an acknowledgment from the network. You should set L3WINDOW to the largest possible value. Note. Some X.25 networks limit the size of L3WINDOW. Consult your vendor for more information. PEXQMSWN and PEXQMSAT Modifiers The disk file $SYSTEM.SYSnn.
PAGE 246
PEXQMNAM Modifiers Configuring Multi-Line Paths Table 13-4. PEXQMSWN and PEXQMSAT Modifiers (page 2 of 2) Modifier Default Value Range of Values INTERFACE_RS422 L2DISCARDONRESET_OFF L2DISCARDONRESET_ON 3 L2RETRIES 10 1 through 255 L2TIMEOUT 100 20 through 32767 LINEPRIORITY 1 1 through 9 LINETF 0 0 through 186 PROGRAM $SYSTEM.CSSnn. C1097P00 (directconnect) $SYSTEM.CSSnn.
PAGE 247
PEXQMIP Modifiers Configuring Multi-Line Paths Table 13-5.
PAGE 248
PEXQMIP Modifiers Configuring Multi-Line Paths Table 13-6.
PAGE 249
PEXQMATM Modifiers Configuring Multi-Line Paths PEXQMATM Modifiers The disk file $SYSTEM.SYSnn.PEXQMATM defines modifiers for Expand-over-ATM lines in multi-line paths. Table 13-6 lists the default value and range of values for each modifier in this profile, if applicable. For modifiers that are mutually exclusive, a check mark (3) is shown in the “Default Value” column to indicate which modifier is present in the profile. Table 13-7.
PAGE 250
Configuring Multi-Line Paths Expand Configuration and Management Manual—529522-002 13 -22 PEXQMATM Modifiers
PAGE 251
Part III.
PAGE 252
Part III.
PAGE 253
14 Subsystem Control Facility (SCF) Commands This section describes the Subsystem Control Facility (SCF) interface to the Expand subsystem and provides SCF command syntax. You should refer to the SCF Reference Manual for G-Series RVUs for general information about running SCF.
PAGE 254
Overview of the Expand Subsystem SCF Interface Subsystem Control Facility (SCF) Commands Overview of the Expand Subsystem SCF Interface The Expand subsystem SCF interface is provided to configure, control, and display information about configured objects within the Expand subsystem.
PAGE 255
Subsystem Control Facility (SCF) Commands • • Expand Subsystem Objects The path function corresponds to the functions defined by Layers 3 and 4 of the Open Systems Interconnection (OSI) Reference Model. You specify the PATH object when you want to display Layer 3 and 4 information or alter Layer 3 and 4 attributes for a single-line Expand line-handler process. The line function corresponds to the functions defined by Layer 2 of the OSI Reference Model.
PAGE 256
Subsystem Control Facility (SCF) Commands Object States The following are some typical device names: $SYS1 An Expand line-handler process that manages a single line to the node named \SYS1 $PATH A path logical device $LINE1, $LINE2, and so on Line logical devices PROCESS Object The PROCESS object type may refer to the Expand manager process ($ZEXP), the network control process ($NCP), or an Expand line-handler process.
PAGE 257
SCF Commands and Objects Subsystem Control Facility (SCF) Commands SCF Commands and Objects Table 14-1 lists the SCF commands and objects that are applicable to the Expand subsystem. Table 14-1.
PAGE 258
Wild-Card Support Subsystem Control Facility (SCF) Commands Table 14-2 lists the sensitive and nonsensitive Expand SCF commands. Table 14-2.
PAGE 259
Subsystem Control Facility (SCF) Commands SCF and the WAN Subsystem SCF and the WAN Subsystem On Integrity NonStop NS-series servers, you use the SCF interface to the WAN subsystem to create $NCP and the Expand line-handler processes. You can also use the SCF interface to the WAN subsystem to perform certain network-management tasks. The SCF interface to the WAN subsystem is described in the WAN Subsystem Configuration and Management Manual.
PAGE 260
Subsystem Control Facility (SCF) Commands • SCF and the SLSA Subsystem Use the Expand subsystem STOP PATH or STOP LINE command before you use the STOP DEVICE command to stop the Expand line-handler process in the primary and backup processors. For a complete comparison of the Expand and WAN subsystem SCF interfaces, refer to Appendix B, Expand and WAN SCF Comparison.
PAGE 261
Subsystem Control Facility (SCF) Commands • Examples You can abort several lines or paths with a single ABORT command by specifying multiple PATH or LINE objects using parentheses as follows: PATH ( path-name , path-name [ , path-name ] ...) LINE ( line-name , line-name [ , line-name ] ...
PAGE 262
Subsystem Control Facility (SCF) Commands ALTER Command ALTER Command The ALTER command changes the values for PATH object types, LINE object types, and the PROCESS $NCP object type. ALTER is a sensitive command. The ALTER command syntax is as follows: ALTER { PROCESS $NCP | PATH path-name | LINE line-name } ALTER DEVICE Command The WAN subsystem ALTER DEVICE command changes the values of a data communications subsystem object.
PAGE 263
Subsystem Control Facility (SCF) Commands ALTER PATH Command ALTER PATH Command The ALTER PATH command is described below. The PATH object type takes the following form: PATH path-name attribute-spec [, attribute-spec ] ...
PAGE 264
Subsystem Control Facility (SCF) Commands Considerations Considerations • You can alter several paths with a single ALTER command by specifying multiple PATH objects using parentheses as follows: -> PATH ( path-name , path-name [ , path-name ] ... ) Examples The following SCF command changes the value of the path’s NEXTSYS attribute to system 100 and the value of its TIMERINACTIVITY attribute to 9 minutes and 30 seconds: -> ALTER PATH $PATH1, NEXTSYS 100, TIMERINACTIVITY 9:30.
PAGE 265
Subsystem Control Facility (SCF) Commands [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ ALTER LINE Command AFTERMAXRETRIES { DOWN | PASSIVE } ] ASSOCIATEDEV device-name ] ASSOCIATESUBDEV subdevice-name ] ATMSEL selector-byte ] CALLTYPE { PVC | SVC | ATMSAP} ] CLBDWNLOADRETRIES integer ] CLBDWNLOADTIMR time ] CLBIDLETIMER time ] CLOCKMODE { DTE | DCE } ] CLOCKSPEED { 600 | 1200 | 2400 | 4800 | 9600 | 19200 | 38400 | 56000 | 115200 } ] CONNECTTYPE { ACTIVEANDPASS
PAGE 266
Subsystem Control Facility (SCF) Commands ALTER LINE Command Table 14-4 lists the line attributes that have corresponding profile modifiers. Table 14-4.
PAGE 267
Subsystem Control Facility (SCF) Commands ALTER LINE Command Table 14-4.
PAGE 268
ALTER LINE Command Subsystem Control Facility (SCF) Commands The valid range for this attribute is 0 to 5:27.67 minutes. The default is 10.00 seconds. DSRTIMER time specifies the amount of time that the line-handler process should wait after a Data Set Ready (DSR) signal from the modem has shut off before it returns a modem status message. This attribute applies to ServerNet wide area network (SWAN) concentrators only. The time interval is specified in the format described in Time Values on page 14-6.
PAGE 269
ALTER LINE Command Subsystem Control Facility (SCF) Commands TIMERBIND time specifies the time interval that the Expand-over-NAM or Expand-over-ServerNet line-handler process will wait for a completion of its bind request to the NAM process. The time interval is specified in the format described in Time Values on page 14-6. The TIMERBIND attribute does not apply to Expand-over-IP and Expand-overATM line-handler processes. A value of 0 indicates an indefinite interval (no timer).
PAGE 270
Considerations Subsystem Control Facility (SCF) Commands Expand-over-X.25 lines: 1 through 32767 Expand-over-SNA lines: 1 through 32767 Expand-over-ServerNet lines: 30 through 32767 The default values for this attribute are as follows: Expand-over-IP lines: 1 Expand-over-ATM lines: 1 Expand-over-X.
PAGE 271
Considerations Subsystem Control Facility (SCF) Commands Table 14-5.
PAGE 272
Examples Subsystem Control Facility (SCF) Commands Table 14-5.
PAGE 273
Subsystem Control Facility (SCF) Commands ALTER PROCESS Command ALTER PROCESS Command The ALTER PROCESS command changes the values of the attributes of the network control process ($NCP). This command changes only the specified attributes of $NCP. ALTER PROCESS is a sensitive command. The ALTER PROCESS command for $NCP has the following syntax: ALTER PROCESS $NCP attribute-spec [ attribute-spec ] ...
PAGE 274
Subsystem Control Facility (SCF) Commands ALTER PROCESS Command AUTOREBALTIME time | ( time, start-time ) determines when automatic rebalancing of multi-CPU paths on the system will occur. When time is specified, rebalancing will occur periodically at the time interval specified starting after the command is executed. When (time, start-time) is specified, rebalancing will occur periodically at the time interval specified starting at the time of day specified in start-time.
PAGE 275
Subsystem Control Facility (SCF) Commands Example MSG48 { ON | OFF } enables (ON) or disables (OFF) the reporting of event message 48 to the EMS collector, $0. Event message 48 is equivalent to console message 48. This is a critical message. It means a change in processor status has occurred at the indicated system. The default value is OFF. MSG49 { ON | OFF } enables (ON) or disables (OFF) the reporting of event message 49 to the EMS collector, $0. Event message 49 is equivalent to console message 49.
PAGE 276
Subsystem Control Facility (SCF) Commands Examples Examples The following SCF command removes from the NRT all the names of systems that are not connected within the network: -> DELETE ENTRY $NCP.* The following SCF command removes the system name \NODEA from the NRT if the system named \NODEA is not connected within the network: -> DELETE ENTRY $NCP.\NODEA INFO Command The INFO command displays the current or default attribute values for the specified objects. INFO is a nonsensitive command.
PAGE 277
INFO PATH Command Subsystem Control Facility (SCF) Commands INFO PATH Command The display for a path without the DETAIL option has the format as shown in Example 14-1. The asterisk (*) indicates that the attribute can be altered using the ALTER command, described earlier in this section. Example 14-1. INFO PATH Command -> INFO PATH $LHPATH EXPAND Info Path Name $LHPATH *Compress ON *Nextsys #255 *L4Retries 3 *L4Timeout 0:00:20.00 Name is the device name of the path.
PAGE 278
Subsystem Control Facility (SCF) Commands INFO PATH Command Example 14-2. INFO PATH, DETAIL Command -> INFO PATH $LHPATH, DETAIL EXPAND Detailed Info PATH $LHPATH *Compress.... ON *OStimeout... 0:00:03.00 *L4Timeout... 0:00:20.00 *L4ExtPackets ON Local *PathBlockBytes 0 *PathPacketBytes 1024 *Nextsys........ #255 *OSspace..... 32767 *L4Retries...... 3 *PathTF.. 1 *L4SendWindow... 254 TimeFactor 1 *L4CongCtrl..... ON *Superpath...
PAGE 279
Subsystem Control Facility (SCF) Commands INFO PATH Command L4Timeout reports the time interval for the Layer 4 timer. L4SendWindow is the maximum number of outstanding packet send requests in any single transport connection. TimeFactor reports the current time factor for this path. The time factor is used by NCP when calculating the best route between systems and represents the cost of using the path. The lower the time factor, the more desirable the path.
PAGE 280
Subsystem Control Facility (SCF) Commands Considerations Superpath reports ON if the path is configured to be a member of a multi-CPU path and OFF if it is not. The Expand line-handler process at the other end of the path must be configured with SUPERPATH_ON or the multi-CPU path feature will not be enabled. You can display the current setting of the SUPERPATH attribute using the STATUS PATH command.
PAGE 281
INFO LINE Command Subsystem Control Facility (SCF) Commands INFO LINE Command The format of the INFO LINE display varies according to the line-handler process type. The first three lines of the display are common for all line types. The rest of the lines vary according to the line type.
PAGE 282
Subsystem Control Facility (SCF) Commands Direct-Connect and Satellite-Connect Line-Handler Processes SpeedK calculates the time factor of the line for the Expand routing algorithm. A value of NOT_SET means that this parameter was not set. See SPEEDK n on page 16-22 for a discussion of SPEEDK. L2Timeout reports the time interval of the Layer 2 T1 timer.
PAGE 283
Subsystem Control Facility (SCF) Commands Direct-Connect and Satellite-Connect Line-Handler Processes Framesize specifies the maximum size frame that can be sent in the network; smaller frames may be sent. The Expand subsystem also uses the FRAMESIZE modifier value to calculate the packet size and determine the size of the frame buffers. If the default FRAMESIZE modifier value is used, the packet size is 132 words. Rsize specifies the time factor of the line for the Expand routing algorithm.
PAGE 284
Subsystem Control Facility (SCF) Commands Direct-Connect and Satellite-Connect Line-Handler Processes in the QualityThreshold before taking the action specified in the parameter DownIfBadQuality. TxWindow reports the number of Expand packets that the line-handler process can send before receiving a reply. Address specifies the Layer 2 primary and secondary addresses. Addresses are system numbers.
PAGE 285
Subsystem Control Facility (SCF) Commands Direct-Connect and Satellite-Connect Line-Handler Processes DRtimeout specifies the time interval that the line-handler process Communications Access Process (CAP) will wait for a response to a request it has sent to the communications line interface processor (CLIP). CLBIdleTimer specifies the time interval between communications line interface processor (CLIP) status probes.
PAGE 286
Expand-Over-IP Line-Handler Processes Subsystem Control Facility (SCF) Commands Program reports the file name of the communications line interface processor (CLIP) program that will be downloaded. LineTF is the line time factor. LINETF has a range of 0 to 186, with a default of 0 (unset). If you set LINETF, it overrides the RSIZE, SPEED, or SPEEDK parameters in calculating the time factor for the line (PATHTF overrides all parameters, including LINETF).
PAGE 287
Expand-Over-IP Line-Handler Processes Subsystem Control Facility (SCF) Commands Aftermaxretries is the line state once retries have been exhausted for the line. DOWN means the line state will be down. PASSIVE means the Expand-over-IP process will issue passive connect requests. Example 14-6 shows the display format for a LINE object with the DETAIL option for Expand-over-IP line-handler processes for IPv4 lines. The asterisk (*) indicates an alterable attribute. Example 14-6.
PAGE 288
Expand-Over-IP Line-Handler Processes Subsystem Control Facility (SCF) Commands Example 14-7. INFO LINE, DETAIL Command, Expand-Over-IP Line-Handler Processes for IPv6 Lines -> INFO LINE $IPTAH0, DETAIL EXPAND Detailed Info LINE $IPTAH0 (LDEV 175) L2Protocol Net^Ip TimeFactor...... 3 Framesize....... 132 -Rsize........... 3 *LinePriority.... 1 StartUp......... OFF *DownIfBadQuality OFF *QualityThreshold 96 *Txwindow........ 7 *Maxreconnects... 0 *Timerreconnect 0:00:30.00 *Retryprobe......
PAGE 289
Subsystem Control Facility (SCF) Commands Expand-Over-IP Line-Handler Processes LinePriority This can be set in the range 1 to 9. The default is 1. The higher the number, the lower priority to use that line. If lines have equal priority, the relative line speeds and transmission delays are used to select the next line. Startup indicates whether the line will be enabled (ON) or disabled (OFF) after a system load. Delay is the expected line time required for a bit to arrive at the other end of the line.
PAGE 290
Subsystem Control Facility (SCF) Commands Expand-Over-IP Line-Handler Processes Timerreconnect is the time interval the Expand-over-IP line-handler process will wait for a successful connection. Retryprobe is the number of times the Expand-over-IP line-handler process will retry the probe of the remote Expand-over-IP line-handler process before concluding that the network is unavailable.
PAGE 291
Subsystem Control Facility (SCF) Commands Expand-Over-ATM Line-Handler Processes SrcIpAddr is the TCP/IP address used by the local Expand-over-IP line-handler process. It is used only if the IPVER is IPv4. SrcIpPort is the port number used by the local Expand-over-IP line-handler process. It is used for both IPVER IPv4 and IPv6. V6DestIpAddr is the destination NonStop TCP/IPv6 address used by the remote Expand-over-IP line-handler process. It is used only if the IPVER is IPv6.
PAGE 292
Expand-Over-ATM Line-Handler Processes Subsystem Control Facility (SCF) Commands Associatedev reports the name of the ATM line associated with the Expand-over-ATM linehandler process. Associatesubdev reports the name of the ATM service access point (SAP). The only currently supported ATM SAP is #IP. Example 14-9 shows the display format for a LINE object with the DETAIL option for Expand-over-ATM line-handler processes that use permanent virtual circuits (PVCs).
PAGE 293
Subsystem Control Facility (SCF) Commands Expand-Over-ATM Line-Handler Processes Rsize specifies the time factor of the line for the Expand routing algorithm. RSIZE can be 0 if the time factor is set using some other modifier. Speed calculates the time factor of the line for the Expand routing algorithm. LinePriority This can be set in the range 1 to 9. The default is 1. The higher the number, the lower priority to use that line.
PAGE 294
Subsystem Control Facility (SCF) Commands Expand-Over-ATM Line-Handler Processes Maxreconnects is the maximum number of times the Expand-over-ATM line-handler process will try to connect to the remote system. AfterMaxRetries is the line state after all retries have been exhausted for the line. Timerreconnect is the time interval the Expand-over-ATM line-handler process will wait for a successful connection.
PAGE 295
Subsystem Control Facility (SCF) Commands Expand-Over-NAM and Expand-Over-ServerNet Line-Handler Processes PvcName is the name of the permanent virtual circuit (PVC). LineTF is the line time factor. LINETF has a range of 0 to 186, with a default of 0 (unset). If you set LINETF, it overrides the RSIZE, SPEED, or SPEEDK parameters in calculating the time factor for the line (PATHTF overrides all parameters, including LINETF).
PAGE 296
Subsystem Control Facility (SCF) Commands Expand-Over-NAM and Expand-Over-ServerNet Line-Handler Processes L2Timeout reports the time interval of the Layer 2 T1 timer. Associatedev reports the name of the X25AM or SNAX/APN process associated with the Expand-over-NAM process. For Expand-over-ServerNet line-handler processes, this field shows $ZZSCL. Associatesubdev reports the name of the NAM subdevice that will be activated by the Expand-overNAM process. The subdevice name for Expand-over-X.
PAGE 297
Subsystem Control Facility (SCF) Commands Expand-Over-NAM and Expand-Over-ServerNet Line-Handler Processes Framesize specifies the maximum size frame that can be sent in the network; smaller frames may be sent. The Expand subsystem also uses the FRAMESIZE modifier value to calculate the packet size. The FRAMESIZE modifier value also determines the size of the frame buffers. Rsize displays the time factor of the line for the Expand routing algorithm.
PAGE 298
Subsystem Control Facility (SCF) Commands Expand-Over-NAM and Expand-Over-ServerNet Line-Handler Processes Maxreconnects is the maximum number of times the Expand-over-NAM or Expand-over-ServerNet line-handler process will try a connect request after successfully binding to the NAM interface. AfterMaxRetries lists the line state once all retries have been exhausted for the Expand-over-NAM or Expand-over-ServerNet line-handler process. DOWN causes the line state to be down.
PAGE 299
Subsystem Control Facility (SCF) Commands Considerations requesting disconnection from the network service provided by the NAM process. See Time Values on page 14-6 for a description of the time interval format. Timerinactivity does not apply to Expand-over-ServerNet line-handler processes. Connecttype lists the Layer 2 Expand-over-NAM or Expand-over-ServerNet line-handler process connect type.
PAGE 300
Subsystem Control Facility (SCF) Commands INFO PROCESS Command CONNECTS displays the systems that are connected or connecting, and only the entry for which the connection is established. If the path is a superpath, the CONNECTS option displays all of the paths in the superpath. LINESET displays the status of a selected path and the status of the started lines that make up that path. Note.
PAGE 301
Subsystem Control Facility (SCF) Commands INFO PROCESS Command sys-a is {\system-name | system-number }. sys-b is {\system-name | system-number }. sys-c is {\system-name | system-number }. If the NETMAP, SUPERPATH, or RPT option is chosen, only one system can be specified. If the SUPERPATH option is chosen, the display lists the multi-CPU paths on a remote node. If the RPT option is chosen, the display lists the reverse pairing table (RPT) on a remote node.
PAGE 302
INFO PROCESS Command Subsystem Control Facility (SCF) Commands DETAIL causes detailed information of $NCP attributes to be displayed. If not specified, only one line of $NCP attribute information is displayed. The display for the INFO PROCESS $NCP command without the DETAIL option has the format as shown in Example 14-12. The asterisk (*) denotes an alterable attribute. Example 14-12.
PAGE 303
INFO PROCESS Command Subsystem Control Facility (SCF) Commands Example 14-13. INFO PROCESS $NCP, DETAIL Command -> INFO PROCESS $NCP, DETAIL EXPAND Detailed Info PROCESS Max System Number.. Algorithm.......... *Connecttime........ *Maxtimeouts........ *NetworkDiameter.... *Message 43......... Message 45......... Message 47......... *Message 49......... Next Rebalance Time Trace File Name.... $NCP AT \NODEA (117) 254 *Aborttimer......... MODIFIEDSPLIT AutomaticMaptimer.. 0:00:00.00 Framesize.........
PAGE 304
Subsystem Control Facility (SCF) Commands INFO PROCESS Command Framesize is used by $NCP to compute the maximum size, in words, of a distance vector (DV) packet. Note. The network control process FRAMESIZE modifier and the Layer 2 SCF FRAMESIZE modifier have the same name. Both the network control process and the Layer 2 FRAMESIZE modifiers are configured using the SCF interface to the WAN subsystem.
PAGE 305
Subsystem Control Facility (SCF) Commands INFO PROCESS Command Message 47 reports whether the reporting of event message 47 to $0 is enabled (ON). Message 47 is a critical message. It means that an end-to-end acknowledgment was not received from the indicated system within the configured Layer 4 timeout interval. Message 48 reports whether the reporting of event message 48 to $0 is enabled (ON) or disabled (OFF). Message 48 is a critical message.
PAGE 306
CONNECTS Option Subsystem Control Facility (SCF) Commands AutoRebalTime reports the time interval for automatic rebalancing of the multi-CPU paths on the system. Rebalancing will occur periodically at the time interval shown. The time interval is displayed in the format described for the Next Rebalance Time attribute. Trace File Name the name of the trace file specified in the SCF TRACE command.
PAGE 307
Subsystem Control Facility (SCF) Commands LINESET Option For multi-CPU paths, the asterisk has a different meaning for non-neighbor nodes than for neighbor nodes. For non-neighbor nodes, the asterisk indicates the Expand line-handler process selected for the pair between the local node and each remote node; all traffic to the remote node uses the indicated line-handler process.
PAGE 308
LINESET Option Subsystem Control Facility (SCF) Commands Example 14-15.
PAGE 309
Subsystem Control Facility (SCF) Commands LINESET Option TF indicates time factors in this display. To use old time-factor values, use the command INFO PROCESS $NCP, OLDLINESET. If you are using the OLDLINESET option on a G06.20 node, the command INFO PROCESS $NCP, LINESET, AT \remote, where \remote is a G06.19 node, displays super time factor information, and the command INFO PROCESS $NCP, OLDLINESET, AT \remote displays non-super time factor information. PID is the process ID.
PAGE 310
NETMAP Option Subsystem Control Facility (SCF) Commands NETMAP Option The display for the INFO PROCESS $NCP command with the NETMAP option has the format as shown in Example 14-16: Example 14-16.
PAGE 311
Subsystem Control Facility (SCF) Commands NETMAP Option #LINESETS=n indicates that there are n communications paths (LINESETS) directly connected to the selected system. The LINESETS are listed in detail after the NETMAP table. The systems in the network are listed by the system number followed by the system name. SYSTEM indicates the number and the name of the system, or node.
PAGE 312
Subsystem Control Facility (SCF) Commands NETMAP Option TF indicates time factors in this display. If you are using the OLDNETMAP option on a G06.20 node, the command INFO PROCESS $NCP, LINESET, AT \remote, where \remote is a G06.19 node, displays super time factor information, and the command INFO PROCESS $NCP, OLDLINESET, AT \remote displays non-super time factor information. PID is the process ID. LINE indicates the device name of a line.
PAGE 313
PATHSET Option Subsystem Control Facility (SCF) Commands PATHSET Option The PATHSET option displays the NCP pathmap information, similar to the LINESET option but in a different format. This format displays both the line-handler LDEV and name, as well as the other information already in the LINESET option. Example 14-17.
PAGE 314
RPT Option Subsystem Control Facility (SCF) Commands Name indicates the device name of the line. Ldev indicates the logical device (LDEV) number associated with each line logical device. Status indicates the status of the line; whether it is ready or not ready. FileErr shows the most recent file system error number, if any, associated with each line. For recovery information on file errors, refer to Identifying Network Problems on page 20-3.
PAGE 315
SUPERPATH Option Subsystem Control Facility (SCF) Commands NEIGHBOR indicates the neighbor node that data is transmitted to over the path. SYS/LDEV indicates the number and the name of the system, or node, and the logical device (LDEV) number. SUPERPATH Option The display for the INFO PROCESS $NCP command with the SUPERPATH option has the format as shown in Example 14-19: Example 14-19.
PAGE 316
SYSTEMS Option Subsystem Control Facility (SCF) Commands LF indicates the load factor for the path in a multi-CPU path (superpath). The effective time factor (ETF) is calculated based on the load factor (ETF = LF * TF). LCPU indicates the local processor number. RCPU indicates the remote processor number. Superpath Rebalancing Considerations A Superpath rebalance can introduce a temporary disruption in the network, similar to but, in general, less than that caused by an Expand path change.
PAGE 317
Subsystem Control Facility (SCF) Commands PRIMARY PROCESS Command system listed in the leftmost column. (Refer to Routing and Time Factors on page 17-21 for more information on the TF.) An asterisk (*) indicates the Expand line-handler process selected for traffic to each known node in the network; this is also the line-handler process used for the $NCP connection protocol with each node. For multi-CPU paths, the asterisk has a different meaning for non-neighbor nodes than for neighbor nodes.
PAGE 318
Subsystem Control Facility (SCF) Commands Considerations cpu-number is the processor number that will now become the primary processor for the specified line or path. Considerations • • • • • If the specified processor is not either the backup or primary processor, an error is returned. If the specified processor is currently the primary processor, a warning is returned. The PRIMARY PROCESS command is not supported directly for Expand-over-IP or Expand-over-PTCPIP line-handler processes.
PAGE 319
Subsystem Control Facility (SCF) Commands PROBE PROCESS Command PROBE PROCESS Command The PROBE PROCESS command applies only to $NCP. PROBE displays the current paths to one or more, or all, of the remote systems within a network, from a specified system within the network. PROBE PROCESS is a nonsensitive command.
PAGE 320
PROBE PROCESS Command Subsystem Control Facility (SCF) Commands Assume that you have entered the following commands: -> SYSTEM \NODEA -> PROBE PROCESS $NCP, TO (\NODEB, \NODEC, \NODED, & -> \NODER, \NODEQ) The display resulting from these commands has the format as shown in Example 14-21: Example 14-21.
PAGE 321
Subsystem Control Facility (SCF) Commands START Command \NODEB. The value in parentheses (00003 ms) indicates that the round-trip time for this probe was 3 milliseconds. 7 \NODEQ - * (00003 ms) indicates that the probe was made from \NODEA to \NODEQ. The list begins with \NODEQ and ends at the system from which the probe was made, indicated by the asterisk *. The connection is direct; there is no system in between.
PAGE 322
Subsystem Control Facility (SCF) Commands Examples Examples The following SCF command starts a line named $LHCMP2: -> START LINE $LHCMP2 The following SCF command starts a path named $PTS and all lines associated with it: -> START PATH $PTS The following SCF commands starts lines named $LHCMP3 and $LHCMP4: -> START LINE ($LHCMP3,$LHCMP4) STATS Command The STATS command displays statistical information about Expand paths and lines and $NCP.
PAGE 323
STATS PATH Command Subsystem Control Facility (SCF) Commands RESET resets the statistical counters for the specified path. This is a sensitive command. The display for a PATH object has the format as shown in Example 14-22: Example 14-22. STATS PATH Command -> STATS PATH $IPTAH1 EXPAND Stats PATH $IPTAH1, PPID ( 2, Reset Time.... FEB 10,2003 14:49:57 Current Ext Mem KBytes Used Number of Known Systems Ext Mem Allocation Fails Current QIO KBytes Used Current QIO MDs Used ------------------------<= 64 .
PAGE 324
Subsystem Control Facility (SCF) Commands STATS PATH Command Reset Time is the last time the statistics counters were reinitialized. Sample Time is the time of the current statistics display. Current Ext Mem KBytes Used is the current amount of extended memory used, in KBytes. Max Ext Mem KBytes Used is the maximum amount, in KBytes, of extended memory used since the last statistics reset or line-handler process start. Number of Known Systems is the total number of nodes known to this path.
PAGE 325
Subsystem Control Facility (SCF) Commands STATS PATH Command Current QIO MDs Used indicates the current QIO message descriptors used. A message descriptor is an internal structure used for sending and receiving messages to and from QIO. Max QIO MDs Used indicates the maximum QIO message descriptors used since the last statistics reset or line-handler process start. A message descriptor is an internal structure used for sending and receiving messages to and from QIO.
PAGE 326
Subsystem Control Facility (SCF) Commands STATS PATH Command Cur OOS in K Bytes is the amount of memory, in kilobytes, currently being used to store packets received out of sequence on this path. Max OOS Used in K Bytes is the maximum amount of memory, in kilobytes, that has been used to store packets received out of sequence on this path.
PAGE 327
Subsystem Control Facility (SCF) Commands STATS PATH Command Looping Packets is the number of incoming packets discarded because they contained the same source ID as the receiver. This can happen if the underlying transport medium is looping back packets or if there is a system with a duplicate node number in the network. Pckt Too Small/Large is the number of incoming packets discarded because they contained either less or more data than expected when the packet was read into the local buffers.
PAGE 328
Subsystem Control Facility (SCF) Commands Considerations ReXmit Timeouts is the number of retransmission timeouts. ReXmit Packets is the number of retransmitted packets. ReIdle Timeouts is the number of idle timeouts causing the congestion window to be reduced. Considerations You can display statistics for several paths with a single STATS PATH command by specifying multiple PATH objects using parentheses as follows: PATH ( path-name , path-name [ , path-name ] ...
PAGE 329
STATS PATH NODE Command Subsystem Control Facility (SCF) Commands node-name is the destination node name, such as \NODEA. RESET resets the statistical counters for the PATH to NODE. This is a sensitive command. The display for a NODE object has the format as shown in Example 14-23: Example 14-23. STATS PATH NODE Command SCF > STATS PATH $SC082,TO \NODEA EXPAND Stats PATH $SC082, PPID ( 2, 299), BPID ( 3, 292) STATS TO NODE \NODEA (82) Reset Time.... AUG 27,2002 14:47:34 Sample Time..
PAGE 330
Subsystem Control Facility (SCF) Commands STATS PATH NODE Command Sample Time is the time of the current statistics display. MESSAGE HISTOGRAM is the overall count of messages sent and received by this node over this path, classified by size in bytes since statistics were last reset using the STATS RESET command, or since the line-handler process was started. The counts do not include any passthrough traffic. Note that not every request is completed, because a CANCEL request might have been issued.
PAGE 331
Subsystem Control Facility (SCF) Commands STATS PATH NODE Command ReXmit Packets is the number of retransmitted packets. ReIdle Timeouts is the number of idle timeouts causing the congestion window to be reduced. Current CWND displays the current congestion control window (CWND) value. Max CWND displays the maximum congestion control window (CWND) value attained. QUEUE DEPTHS displays the queue depth statistics for the specified path. Pending displays the number of pending requests queued.
PAGE 332
Examples Subsystem Control Facility (SCF) Commands Examples The following SCF command displays statistical information for a path named $PATH1: -> STATS PATH NODE1, TO NODE2 STATS LINE Command The STATS LINE command has the following syntax: STATS [ / OUT file-spec / ] LINE line-name [ , RESET ] / OUT file-spec / causes any SCF output generated by the command to be directed to the specified file. RESET resets the statistical counters for the specified line. STATS LINE is a sensitive command.
PAGE 333
Subsystem Control Facility (SCF) Commands Expand-Over-IP Line-Handler Processes Resettime is the last time the statistics counters were reinitialized. Sampletime is the last time the statistics were collected. Conn Cmd is the command used to initiate a connect with a remote system. A connect command is similar to the HDLC SABM frame. Conn Resp is the response to a connect command. This command completes the lowest level of Expand-over-IP connection establishment.
PAGE 334
Expand-Over-ATM Line-Handler Processes Subsystem Control Facility (SCF) Commands Mem Low is the number of times a memory low indication was given to the Expand-over-IP line-handler process from QIO. If the number is increasing, then the QIO resources are running low. Line Quality is the line-quality value computed every 500 frames. This value is not alterable.
PAGE 335
Subsystem Control Facility (SCF) Commands Expand-Over-ATM Line-Handler Processes Conn Cmd is the command used to initiate a connect with a remote system. A connect command is similar to the HDLC SABM frame. Conn Resp is the response to a connect command. This command completes the lowest level of Expand-over-ATM connection establishment. Data is the number of data frames sent and received. Query Cmd is the command used to probe the system for “I’m alive” status.
PAGE 336
Expand-Over-ServerNet, Expand-Over-X.25, and Expand-Over-SNA Line-Handler Processes Subsystem Control Facility (SCF) Commands Line Quality is the line-quality value computed every 500 frames. This value is not alterable. Line quality is computed using the following formula: 100 * (( TOTAL FRAMES - ERROR FRAMES ) / TOTAL FRAMES) Line Quality reports a value below 100 only when the result of the formula is 95 or less; that is, when less than 95 percent of the packets are error-free.
PAGE 337
Subsystem Control Facility (SCF) Commands Expand-Over-ServerNet, Expand-Over-X.25, and Expand-Over-SNA Line-Handler Processes Sampletime is the last time the statistics were collected. Bind indicates the line handler bind to an associate device (such as $ZZSCL or $X25AM). Aconn indicates the number of connects while in active mode. Pconn indicates the number of connects while in passive mode. An active connect message is expected as the reply.
PAGE 338
SWAN Concentrator Lines Subsystem Control Facility (SCF) Commands device notifies the linehandler of any process changes on the remote system or the connection (such as if the phandle changes). Data indicates the number of data frames received. Normal ServerNet traffic is not counted here because normal data traffic by-passes the line handler for these processes. Proc lookup failures process lookup failures indicate the number of failures to see the associate device.
PAGE 339
Subsystem Control Facility (SCF) Commands SWAN Concentrator Lines BPID is the backup process ID. Resettime is the last timestamp that the statistics counters were reinitialized. Sampletime is the last timestamp that the statistics were collected. LEVEL 2 shows the counts of the Layer 2 frames sent and received by this input-output process (IOP) since statistics were last reset using the STATS RESET command or since the line-handler process was started.
PAGE 340
Subsystem Control Facility (SCF) Commands SWAN Concentrator Lines LEVEL 2 DETAIL is the number of Expand frames sent and received through this input-output process (IOP), shown by frame type. If your system receives a large number of SABM, DISC, RR, or I-FRM(P) frames relative to the total number of information frames (I-Frames), your system might have a noisy line. Refer to Section 20, Troubleshooting, for information on troubleshooting Layer 2 problems. SAMB specifies set asynchronous balanced mode.
PAGE 341
Subsystem Control Facility (SCF) Commands SWAN Concentrator Lines DRIVER displays the counters used to account for errors in received frames. The driver counters apply only to the link between the input-output process (IOP) and the communications line interface processor (CLIP), that is, the CLB. Total Frms is the total number of frames that have been transmitted and received between the communications access process (CAP) and the CLIP since THRESHOLD number of frames was last transmitted.
PAGE 342
Subsystem Control Facility (SCF) Commands SWAN Concentrator Lines Rcv OverRun is the number of frames received that were longer than the maximum frame size expected since statistics were last reset using the STATS RESET command or since the line-handler process was started. This problem is caused by the loss of modem synchronization.
PAGE 343
Subsystem Control Facility (SCF) Commands Considerations Considerations You can display statistics for several lines with a single STATS LINE command by specifying multiple LINE objects using parentheses as follows: LINE ( line-name , line-name [ , line-name ] ...
PAGE 344
Subsystem Control Facility (SCF) Commands STATS PROCESS Command AT { system-list | * } where system-list is ( [ sys-a [ , sys-b [ , sys-c [ , .... ]]]] ). sys-a is { \system-name | system-number }. sys-b is { \system-name | system-number }. sys-c is { \system-name | system-number }. If the NETFLOW option is chosen, only one system name or number can be specified. If the AT option is omitted, the SCF target system name is used.
PAGE 345
STATS PROCESS Command Subsystem Control Facility (SCF) Commands Similarly, if * is specified and the NETFLOW option is specified, the status of the entire network is displayed, as viewed from the system specified in the AT parameter. Assume that you have entered the following command: -> STATS PROCESS $NCP, NETFLOW, AT \N1, TO ( 2, 4, 5, 6, 7, 9, 10, 13, 14, 15) The resulting display has the format as shown in Example 14-28 (example display of $NCP statistics with NETFLOW option): Example 14-28.
PAGE 346
STATS PROCESS Command Subsystem Control Facility (SCF) Commands TOTAL PKTS-SENT reports the total number of packets sent from this system to a selected system since the line-handler process was started. TOTAL LINKS-RCVD reports the total number of link requests received by this system from a selected system since the line-handler process was started. TOTAL PKTS-RCVD reports the total number of packets received by this system from a selected system since the line-handler process was started.
PAGE 347
STATUS Command Subsystem Control Facility (SCF) Commands TOTAL PKTS-RCVD reports the total number of packets received by this system. TOTAL PASSTRU-SENT reports the total number of passthrough packets forwarded from this system. TOTAL PASSTRU-RCVD reports the total number of passthrough packets received by this system. STATUS Command The STATUS command displays the dynamic state, last error, and modifiable values of the specified object. It also displays specific subsystem attributes and values.
PAGE 348
STATUS PATH Command Subsystem Control Facility (SCF) Commands Name is the device name of the path. State indicates the summary state of the path. The path is in the STARTED, STARTING, DIAGNOSING (for SWAN concentrators only), or STOPPED state. PPID is the primary process ID. BPID is the backup process ID. Lines # reports the total number of lines associated with the path. The display for a path with the DETAIL option has the format as shown in Example 14-31: Example 14-31.
PAGE 349
Considerations Subsystem Control Facility (SCF) Commands Trace Status indicates whether the path is being traced. Superpath reports ON if the path is currently a member of a multi-CPU path and OFF if it is not. The Expand line-handler process at the other end of the path must be configured with SUPERPATH_ON or the multi-CPU path feature will not be enabled. The configured value can be displayed using the INFO PATH command.
PAGE 350
STATUS LINE Command Subsystem Control Facility (SCF) Commands Name is the device name of the line. State indicates the summary state of the line. The line is in either the STARTED or STOPPED state. STATUS indicates the status of the line: ready or not ready. PPID is the primary process ID. BPID is the backup process ID. CIU-Path indicates which ServerNet wide area network (SWAN) concentrator path (A or B) is being used by this line to communicate with the SWAN concentrator.
PAGE 351
Subsystem Control Facility (SCF) Commands STATUS LINE Command PPID is the primary process ID. BPID is the backup process ID. State indicates the summary state of the line. The line is in either the STARTED or STOPPED state. Path LDEV contains the logical device (LDEV) number of the path associated with this line. Trace Status indicates whether the line is being traced.
PAGE 352
Subsystem Control Facility (SCF) Commands STATUS LINE Command IP Address is the Internet Protocol (IP) address associated with the SWAN concentrator path (A or B) being used by this line to communicate with the SWAN concentrator. Each SWAN path is assigned a unique IP address. Effective line priority indicates the effective priority of the line. Trace File Name the name of the trace file specified in the SCF TRACE command.
PAGE 353
Subsystem Control Facility (SCF) Commands STATUS LINE Command Detailed State indicates a more detailed state. The detailed states are as follows: ACCEPT indicates that a switched virtual circuit (SVC) connection has been accepted from the remote system. This state applies to Expand-over-ATM line-handler processes that use SVC connections only.
PAGE 354
Subsystem Control Facility (SCF) Commands STATUS LINE Command INACTIVE indicates that the Expand line-handler process is inactive. It is either waiting for data to send or is waiting for an active connect from the other side. LISTEN indicates that the Expand-over-ATM line-handler process is waiting for switched virtual circuit (SVC) connection establishment from the remote system. This state applies to Expand-over-ATM line-handler processes that use SVC connections only.
PAGE 355
STATUS LINE Command Subsystem Control Facility (SCF) Commands SOCKET SETUP an internal state that should not persist. This state applies to Expand-over-IP line-handler processes only. SOCKET_SPACE an internal state that should not persist. This state applies to Expand-over-IP line-handler processes only. WAIT indicates that the Expand line-handler process is waiting for another process or subsystem. Refer to the Detailed Info field for more information.
PAGE 356
Considerations Subsystem Control Facility (SCF) Commands Table 14-7.
PAGE 357
Subsystem Control Facility (SCF) Commands STOP Command STOP Command The STOP command terminates the activity of an object normally. It nondisruptively deletes all connections to and from an object. Upon successful completion, configured objects are left in the STOPPED state and nonconfigured objects are deleted. This is a sensitive command.
PAGE 358
Subsystem Control Facility (SCF) Commands TRACE Command TRACE Command The TRACE command can request the capture of target-defined data items, alter trace parameters, and end tracing. TRACE is a sensitive command. An SCF trace produces a trace file that can be displayed using the commands available in the PTrace program. The trace file is created by SCF. The PTrace program is described in the PTrace Reference Manual and in Section 15, Tracing.
PAGE 359
Subsystem Control Facility (SCF) Commands TRACE Command The TRACE command has the following syntax for tracing the network control process ($NCP): TRACE [ [ [ [ [ [ [ [ / , , , , , , , OUT file-spec / ] PROCESS $NCP BACKUP ] COUNT count ] NOCOLL] PAGES pages ] RECSIZE size] SELECT select-spec ] TO file-spec ] or TRACE PROCESS $NCP , STOP / OUT file-spec / causes any SCF output generated for the command to be directed to the specified file. PATH path-name is the device name of the path to be traced.
PAGE 360
TRACE Command Subsystem Control Facility (SCF) Commands PAGES pages pages is an integer in the range 4 to 64. PAGES controls how much space, in units of pages, is allocated in the extended data segment used for tracing. PAGES may be specified only when the trace is being initiated. The default value is 64 pages. RECSIZE size size is an integer in the range 16 to 4050. It controls the length of the data in the trace data records. The trace header is not included in RECSIZE. The default is 120 bytes.
PAGE 361
TRACE Command Subsystem Control Facility (SCF) Commands Table 14-9.
PAGE 362
Subsystem Control Facility (SCF) Commands Considerations TO file-spec file-spec specifies the file to which tracing is to be initiated. The file may have been previously created by you as an unstructured file with file code 0. WRAP causes the trace segment data to wrap instead of stopping the trace when it reaches the end of file. The default is FALSE. STOP discontinues the trace currently in progress.
PAGE 363
Subsystem Control Facility (SCF) Commands VERSION Command The following SCF command initiates a trace of two lines named $LINE2 and $LINE3: -> TRACE LINE ($LINE2,$LINE3) VERSION Command The VERSION command displays the version level of the Expand manager process ($ZEXP), the network control process ($NCP), or an Expand line-handler process. VERSION is a nonsensitive command.
PAGE 364
Subsystem Control Facility (SCF) Commands VERSION PROCESS Command VERSION PROCESS Command Example 14-35 shows the displays for the VERSION PROCESS command: Example 14-35. VERSION PROCESS Command -> VERSION PROCESS $SC254, DETAIL Detailed VERSION PROCESS \DRP25.
PAGE 365
15 Tracing This section describes the tracing process when the SCF TRACE command is used with commands available in the PTrace facility. The SCF TRACE command allows you to select the records that you want written to a disk file. PTrace commands allow you to select which of those records you want formatted and sent to an output device. The output device can be a terminal, spooler, or printer.
PAGE 366
Why Tracing Is Important Tracing Why Tracing Is Important Tracing allows HP personnel to see the history of a data communications link, including significant points in the internal processing of the traced entity. Isolating a data communications problem using an Expand trace is easier than using a system dump. How to Use Tracing For tracing to be effective, make sure you follow these guidelines: • • • • Always trace both ends of a path.
PAGE 367
Tracing a Line in a Multi-Line Path Tracing Tracing a Line in a Multi-Line Path To start a trace of a line that is part of a multi-line path, enter -> TRACE LINE $line-name, TO $file-name, SELECT ALL, WRAP To stop the trace, enter -> TRACE LINE $line-name, STOP $line-name specifies the name of the line logical device. $file-name specifies the name of the file to which the trace records will be written.
PAGE 368
Tracing Using SCF Tracing Figure 15-1 shows the relationship of the tracing process components when SCF is used. Figure 15-1.
PAGE 369
PTrace Command Overview Tracing PTrace Command Overview When you are using the PTrace facility, consider the following: • • • You have not been provided trace-format information to read these formats because you do not have the source code. Therefore, when reporting problems, select the ALL option available in the SCF TRACE command. You should always specify the source disk file using the PTrace FROM command before any other PTrace command.
PAGE 370
FILTER Command Tracing FILTER Command The FILTER command prevents the selected type of information from being sent to the output device. FILTER { option | option,option,...option | RESET } option defines the type of information you do not want to display or print to the output device. You can specify one or more options separated by commas: NOHDR filters trace record header information. NOL2 filters Layer 2 frame header information. NOL2RR filters Layer 2 Receive Ready (RR) frame information.
PAGE 371
Examples Tracing • Issue the FILTER command with the RESET option. If you issue the FILTER command with one set of selection options and then reissue it with a different set of selection options, the options entered with the second FILTER command are used to determine the trace information sent to the output device. The previously entered selection options are overridden; selection options are not cumulative.
PAGE 372
Examples Tracing record in which the previously specified string parameter was found. However, if you enter the FIND command without a string parameter and no previous FIND command with a string parameter has been issued, an error is returned. While the PTrace facility processes the FIND command, trace records will not be sent to the output device. If the specified string is found in an output line, the entire record is sent to the output device.
PAGE 373
Example Tracing Example ?HEX ON LABEL Command The LABEL command formats state machine entries, frames, packets, message headers, and message data when set to ON (or defaults). This command is useful only for personnel who have source code listings. LABEL { ON | OFF } ON | OFF ON enables formatting of trace record information. This is the default when first entering PTrace. OFF disables formatting of trace record information.
PAGE 374
Example Tracing Table 15-2. Number of Trace Lines Displayed F Key Number of Lines F Key Number of Lines F1 1 F9 9 F2 2 F10 10 F3 3 F11 11 F4 4 F12 12 F5 5 F13 13 F6 6 F14 14 F7 7 F15 15 F8 8 F16 16 Example ?NEXT 15 AFTER 13:01 OCTAL Command The OCTAL command, when set to ON, prints the data portion of a trace record, including the record header, in octal format. OCTAL { ON | OFF } ON | OFF ON enables printing in octal format.
PAGE 375
Example Tracing STOP closes the spooler or line printer specified in the previous OUT command. As a result, subsequent trace records are displayed at your terminal. Example ?OUT $s.#tester RECORD Command The RECORD command displays selected records by number. You can select records individually, in a range, or ALL. If you select records within a range, only records or record portions that meet the criteria you have defined using the SELECT and FILTER commands are displayed.
PAGE 376
SELECT Command Tracing SELECT Command The SELECT command sets the selection criteria for the record types sent to the output device. When PTrace is determining which records to display in response to a NEXT, FIND, or RECORD command, it checks the selection bit mask to determine whether the record is of a type you want to display. This selection criteria is in addition to the selection criteria you have set using the FILTER command. If you do not specify a mask or keyword, ALL bits are set.
PAGE 377
SELECT Command Tracing Hex Mask Octal Mask $NCP SCF Bit PATH Keyword Line Table 15-3.
PAGE 378
SELECT Command Tracing Expand Configuration and Management Manual—529522-002 15 -14
PAGE 379
Part IV.
PAGE 380
Part IV.
PAGE 381
16 Expand Modifiers The Expand subsystem provides many modifiers to allow you to customize your network. These modifiers are contained in the profiles. Some modifiers are required, some are optional, some only appear in certain profiles, and others appear in several profiles. This section describes the modifiers that are related to the configuration of Expand line-handler processes.
PAGE 382
Required Modifiers Expand Modifiers Table 16-1. Required Modifiers (page 2 of 3) Modifier Description ASSOCIATESUBDEV Must be used to specify • • • The name of the X25AM subdevice to which an Expandover-X.25 line-handler process will bind. The subdevice name of the SNAX/APN logical unit (LU) used by an Expand-over-SNA line-handler process. The name of the Asynchronous Transfer Mode (ATM) service access point (SAP) used by an Expand-over-ATM line-handler process.
PAGE 383
Required Modifiers Expand Modifiers Table 16-1. Required Modifiers (page 3 of 3) Modifier Description DESTIPADDR Specifies the Internet Protocol (IP) address used by a remote (destination) Expand-over-IP line-handler process. Required by: Expand-over-IP line-handler processes if IPVER is IPv4. Default: 0.0.0.0. DESTIPPORT Specifies the port number used by a remote (destination) Expand-over-IP line-handler process. Required by: Expand-over-IP line-handler processes if IPVER is IPv4. Default: 1024.
PAGE 384
Modifier Dictionary Expand Modifiers Modifier Dictionary This subsection lists in alphabetical order all the modifiers used to configure Expand line-handler processes and describes each modifier in detail. Default values and value ranges are described, if applicable.
PAGE 385
ASSOCIATESUBDEV #n Expand Modifiers ASSOCIATESUBDEV #n Default: Units: Range: No default for Expand-over-NAM line-handler processes #IP for Expand-over-ATM line-handler processes Not applicable Not applicable This modifier is required for Expand-over-NAM and Expand-over-ATM line-handler processes only. n may specify the following: • • • The name of an X25AM subdevice to which the Expand-over-X.25 line-handler process will bind.
PAGE 386
CLOCKMODE_DCE/CLOCKMODE_DTE Expand Modifiers CLOCKMODE_DCE/CLOCKMODE_DTE Default: Units: Range: CLOCKMODE_DCE Not applicable Not applicable These modifiers are applicable to direct-connect and satellite-connect Expand linehandler processes only. The CLOCKMODE_DCE modifier disables the communications line interface processor (CLIP) clock on the ServerNet wide area network (SWAN) concentrator used by the line. It causes the SWAN concentrator to provide no clocking.
PAGE 387
CONNECTTYPE_ACTIVEANDPASSIVE/ CONNECTTYPE_PASSIVE Expand Modifiers To determine if data compression should be set, you should examine the message traffic character composition to assess its compressibility. For example, EDIT files do not compress well, while structured files and object code compress an average of 20 to 50 percent. If compressed data is received by an Expand line-handler process that does not have compression configured, the data will still be decompressed.
PAGE 388
DESTATMADDR n Expand Modifiers that a data bit spends on the line during message transmission. The Expand linehandler process uses the transmission size, the amount of delay before the message can be dispatched, and the DELAY modifier value to select the most efficient line for data transmission within a path that consists of multiple lines. Transmission delay is usually minimal on a terrestrial link. Transmission delay on a satellite link is greater than on a terrestrial link.
PAGE 389
DESTIPPORT n Expand Modifiers DESTIPPORT n Default: Units: Range: 1024 Not applicable 0 through 65534 This modifier is applicable to Expand-over-IP line-handler processes only. This modifier specifies the port number used by the remote (destination) Expand-over-IP line-handler process. It is the port number specified in the remote line-handler process’ SRCIPPORT modifier. Port numbers are explained in the TCP/IP Configuration and Management Manual.
PAGE 390
FRAMESIZE n Expand Modifiers FRAMESIZE n Default: Units: Range: 132 Words 64 through 2047 This Layer 2 modifier is applicable to all Expand line types. This modifier specifies the maximum size frame that can be sent in the network; smaller frames may be sent. The FRAMESIZE modifier is also used by the Expand subsystem to calculate the packet size, which determines the size of the frame buffers.
PAGE 391
L2DISCARDONRESET_OFF/L2DISCARDONRESE T_ON Expand Modifiers L2DISCARDONRESET_OFF/L2DISCARDONRESET_ON Default: Units: Range: L2DISCARDONRESET_ON Not applicable ON or OFF These Layer 2 modifiers are applicable to direct-connect and satellite-connect linehandler processes only.
PAGE 392
L2TIMEOUT n Expand Modifiers L2TIMEOUT n Default: Units: Range: 100 (1.00 second) for direct-connect lines 200 (2.00 seconds) for satellite-connect lines 0.01 seconds 20 through 32767 This Layer 2 modifier is applicable to direct-connect and satellite-connect line-handler processes only. This modifier specifies the length of time, in one-hundredth of a second increments, that the Expand line-handler process will wait for a response to a request at Layer 2 before retrying.
PAGE 393
L4CONGCTRL_OFF/L4CONGCTRL_ON Expand Modifiers L4CONGCTRL_OFF/L4CONGCTRL_ON Default: Units: Range: L4CONGCTRL_ON for Expand-over-IP and Expand-over-ATM lines L4CONGCTRL_OFF for other line types and multi-line paths Not applicable ON or OFF These path modifiers are applicable to all Expand line types. The L4CONGCTRL_ON modifier enables the congestion control mechanism on the Expand node for sending packets on a path.
PAGE 394
L4SENDWINDOW n Expand Modifiers request before reporting an error. You should read the description of Layer 4 retries in Section 17, Subsystem Description, before using this modifier. Note. The L4RETRIES modifier value should be set to the same value for every Expand linehandler process on every node in the network. L4SENDWINDOW n Default: Units: Range: 254 Packets 187 through 254 This path modifier is applicable to all Expand line types.
PAGE 395
LIFNAME n Expand Modifiers l2timeout is the time interval, in one-hundredth of a second increments, that the Expand line-handler process will wait for a response to a request at Layer 2 before retrying. (You can modify this value using the L2TIMEOUT modifier as described L2TIMEOUT n on page 16-12.) q is the hop count (HC) of the longest end-to-end route in the network. For Expand-over-X.
PAGE 396
MAXRECONNECTS n Expand Modifiers handler will display a configuration error. Refer to Setting Time Factors on page 17-22 for more detailed information about establishing time factors. MAXRECONNECTS n Default: Units: Range: 0 Not applicable 0 through 32767 This modifier is applicable to Expand-over-NAM, Expand-over-IP, Expand-over-ATM, and Expand-over-ServerNet line-handler processes.
PAGE 397
OSTIMEOUT n Expand Modifiers multi-line paths may require all Expand line-handler processes to have more than the default buffer size for storing OOS packets. Note. The OSSPACE modifier is currently ignored. The amount of space allocated to out-ofsequence (OOS) packets is limited by the OSTIMEOUT modifier and by the base size of the line handler’s data segment. The OSSPACE modifier may be used in the future, in which case its default, units, and range may be different.
PAGE 398
PATHPACKETBYTES n Expand Modifiers than the PATHPACKETBYTES modifier value, the PATHBLOCKBYTES modifier value is automatically changed to the PATHPACKETBYTES modifier value. A value of 0 (the default) specifies that the multipacket frame feature will be disabled. A value of 0 is recommended for Expand-over-ATM lines.
PAGE 399
PATHTF n Expand Modifiers PATHTF n Default: Units: Range: 0 (unset) Not applicable 0 through 186 The PATHTF has a range of 0 to 186 to designate the time factor in selecting the best path to other nodes. A smaller number indicates a more desirable path for routing. If you set PATHTF, it overrides any other parameter (RSIZE, SPEED, SPEEDK, or LINETF) in calculating the time factor for the path.
PAGE 400
QUALITYTHRESHOLD n Expand Modifiers QUALITYTHRESHOLD n Default: Units: Range: 0 Integers 0 through 99 If the line reports quality lower than this percentage value, a timer is started. See also DOWNIFBADQUALITY ON/ DOWNIFBADQUALITY OFF and QUALITYTIMER n. This modifier is applicable to both single-line and multi-line IP, ATM, satellite, and SWAN line-handler processes.
PAGE 401
RSIZE n Expand Modifiers RSIZE n Default: None Units: Not applicable Range: 0 through 186 This required modifier specifies the time factor of the line for the Expand routing algorithm. RSIZE must always be set to 1 for $NCP and set to 0 for the path device of a multi-line path. Starting with G06.20, you can use the new parameters, LINETF n and PATHTF n, to set values for lines that will override all other parameters in calculating time factors.
PAGE 402
SPEEDK n Expand Modifiers factors. PATHTF overrides RSIZE, SPEED, SPEEDK, or LINETF, whereas LINETF overrides RSIZE, SPEED, and SPEEDK (but not PATHTF). Either RSIZE, SPEED, SPEEDK, LINETF, or PATHTF must be set, else the line handler will display a configuration error. The formula to convert from SPEED to LINETF is: LINETF = (224000 + (SPEED / 2)) / SPEED Refer to Setting Time Factors on page 17-22 for more detailed information about establishing time factors. SPEEDK n Default: Units: Range: none.
PAGE 403
SRCIPADDR n Expand Modifiers Table 16-2.
PAGE 404
SRCIPPORT n Expand Modifiers This modifier is applicable to Expand-over-IP line-handler processes only. This modifier specifies the Internet Protocol (IP) address associated with the NonStop TCP/IP process used by the local Expand-over-IP line-handler process. Because a NonStop TCP/IP process can have more than one IP address, you must specify to the Expand-over-IP line-handler process which IP address to use. The address must be specified by number (for example, 130.252.12.3).
PAGE 405
TIMERINACTIVITY n Expand Modifiers be no more than 16 multi-CPU paths in a system and each multi-CPU path can consist of no more than 16 paths. Expand line-handler processes at both ends of the path must be configured with SUPERPATH_ON or the multi-CPU feature is not enabled. Expand line-handler processes that use the SUPERPATH_ON modifier also should use congestion control. The extended packet format is required for Expand linehandler processes that are part of a multi-CPU path.
PAGE 406
TIMERRECONNECT n Expand Modifiers commands. For the ADD DEVICE command, the format of the time is just plain seconds. Probes will continue to be sent out the number of times specified by the RETRYPROBE attribute. If the TIMERPROBE/RETRYPROBE cycle expires without a returned status, the Expand-over-NAM, Expand-over-ServerNet, Expand-over-ATM, or Expand-over-IP line-handler process declares the network unavailable. See also RETRYPROBE n and TIMERRECONNECT n.
PAGE 407
V6DESTIPADDR n Expand Modifiers processes, you do not need to have a large TXWINDOW modifier value if the PATHBLOCKBYTES or PATHPACKETBYTES modifier value is large. The product of the TXWINDOW modifier value multiplied by the larger of the PATHPACKETBYTES or PATHBLOCKBYTES modifier values must allow the space for line buffers to fit within 131064 words.
PAGE 408
Profiles Expand Modifiers The default must be changed before the line is started. The address must be specified by number (for example, 31CA:B145:5489:1034:1784:B245:4029:1257). Configuring NonStop TCP/IPv6 addresses is explained in the TCP/IPv6 Configuration and Management Manual. Profiles This subsection lists the modifiers that are contained in each of the profiles.
PAGE 409
Single-Line Expand Line-Handler Process Modifiers Expand Modifiers Table 16-3.
PAGE 410
Multi-Line Path Modifiers Expand Modifiers Table 16-3.
PAGE 411
Multi-Line Path Modifiers Expand Modifiers • • • • • • • • • • • • L4EXTPACKETS_ON L4RETRIES L4SENDWINDOW L4TIMEOUT NEXTSYS OSSPACE OSTIMEOUT PATHBLOCKBYTES PATHPACKETBYTES PATHTF SUPERPATH_OFF SUPERPATH_ON Table 16-4 lists the modifiers in the profiles provided for line-logical devices. Table 16-4.
PAGE 412
Multi-Line Path Modifiers Expand Modifiers Table 16-4.
PAGE 413
Multi-Line Path Modifiers Expand Modifiers Table 16-4.
PAGE 414
Expand Modifiers Multi-Line Path Modifiers Expand Configuration and Management Manual—529522-002 16 -34
PAGE 415
17 Subsystem Description This section provides a high-level technical description of the architecture and dynamics of the Expand subsystem. You should be familiar with the information presented in this section before you attempt to configure, manage, or troubleshoot the Expand subsystem.
PAGE 416
Expand Subsystem Components Subsystem Description Expand Subsystem Components The Expand subsystem comprises the following major components: • • • Expand Line-Handler Processes on page 17-2 Network Control Process ($NCP) on page 17-6 Expand Manager Process ($ZEXP) on page 17-7 Expand Line-Handler Processes An Expand line-handler process is responsible for • • • • Maintaining the communications path between two adjacent nodes.
PAGE 417
Expand Line-Handler Processes Subsystem Description A multi-CPU path is created by associating Expand line-handler processes with one another using the SUPERPATH_ON modifier. Each line-handler process that is a member of a multi-CPU path is configured in a different processor. Note. The path and line functions of an Expand line-handler process are described in more detail in Expand Subsystem and the OSI Reference Model on page 17-8.
PAGE 418
Expand Line-Handler Processes Subsystem Description Expand-Over-NAM Line-Handler Processes Expand-over-NAM line-handler processes use the NETNAM protocol to access the network access method (NAM) interface provided by an X25AM or a SNAX/APN line-handler process. Note. For more information about the Expand-to-NAM interface, refer to Expand-to-NAM Interface on page 17-48. Expand-Over-NAM With X25AM The X25AM subsystem provides access to X.25 packet-switched data networks (PSDNs).
PAGE 419
Expand Line-Handler Processes Subsystem Description configured to use a particular SNAX/APN line and logical unit (LU). At least one SNAX/APN line and one Expand line must be configured and started at each end of the SNA network through which the Expand-over-SNA line-handler processes will communicate. Expand-Over-IP Line-Handler Process The Expand-over-IP line-handler process uses the NonStop TCP/IP subsystem to provide connectivity to an Internet Protocol (IP) network.
PAGE 420
Network Control Process ($NCP) Subsystem Description incoming and outgoing data, usually bypass the Expand-over-ServerNet line-handler process and are handled directly by the ServerNet fabrics and the NonStop cluster switches; the Expand software is not involved. Expand-Over-ATM Line-Handler Process The Expand-over-ATM line-handler process uses the Asynchronous Transfer Mode (ATM) subsystem to provide connectivity to an ATM network.
PAGE 421
Expand Manager Process ($ZEXP) Subsystem Description Network Utility Process Functions The network utility process, $ZNUP, answers requests that must wait for system information. It also responds to requests for the time at remote systems, the process information of remote processes, device-information requests, and traffic statistics. The network utility process runs as logical device number 4.
PAGE 422
Expand Subsystem and the OSI Reference Model Subsystem Description Expand Subsystem and the OSI Reference Model The Expand line-handler process and $NCP components of the Expand subsystem contain some of the functions defined in the lower five layers of the OSI Reference Model. The Expand subsystem does not provide any Application Layer or Presentation Layer functions; these functions, in addition to some Session Layer functions, are provided by the message and file systems.
PAGE 423
Expand Line-Handler Process Layer Functions Subsystem Description Expand Line-Handler Process Layer Functions An Expand line-handler process implements several different protocols, including the HP proprietary End-to-End protocol. These protocols provide some of the functions defined by the lower five layers of the OSI Reference Model. OSI Session Layer (Layer 5) The OSI Session Layer coordinates processes and is responsible for the setup and termination of a communications path.
PAGE 424
Expand Line-Handler Process Layer Functions Subsystem Description OSI Data Link Layer (Layer 2) The OSI Data Link Layer defines the rules for transmission on the physical medium.
PAGE 425
$NCP Layer Functions Subsystem Description $NCP Layer Functions As shown in Figure 17-2, $NCP provides some functions of both the OSI Transport and Network Layers. $NCP at the OSI Transport Layer $NCP provides part of the OSI Transport Layer function because it monitors processor UP and DOWN notifications.
PAGE 426
Path Function of the Expand Subsystem Subsystem Description Path Function of the Expand Subsystem This subsection describes the end-to-end (Layer 3) and packet routing (Layer 4) messages that are generated by the End-to-End protocol. Layers 3 and 4 of the Endto-End protocol provide the path function of the Expand subsystem.
PAGE 427
Protocol Packet Types Subsystem Description Connection Reset (CONN RST) A CONN RST is a connection-establishment–reset-setup packet. This packet is sent by $NCP at one of the two end nodes if a packet sequence problem is detected during connection establishment. Node Status (NODE STAT) A NODE STAT is a connection-establishment–system-status setup packet.
PAGE 428
Protocol Packet Types Subsystem Description Link Cancel Request (LCAN) An LCAN is a control packet that is sent in response to a user request to abort a prior LRQ. Data Packet Acknowledgment (ACK) An ACK is a control packet. It is a positive acknowledgment of a data packet (either an LRQ or an LCMP). LRQ and LCMP packets can include acknowledgments. An ACK is only used to acknowledge data packets if no other data packets are ready to be sent.
PAGE 429
Packet Synchronization Subsystem Description Trace Request (TRACE) A TRACE is a data packet that is sent in response to an SCF PROBE command. It contains the identifier of each node it encounters on its route from its sender to its receiver. PING Message A PING message is sent by an Expand line-handler process to measure the round trip time to a neighbor node.
PAGE 430
Example of End-to-End Protocol Packet Exchanges Subsystem Description Normal Data Exchange Figure 17-3 is an example of an error-free exchange of data. Node \A sends two LRQs to node \B. Node \B sends ACK sequence number 2 to indicate the positive acknowledgment of node \A’s LRQs and then replies to each LRQ with an LCMP. Node \A acknowledges node \B’s LCMPs by sending ACK sequence number 2. Note. The sequence number of an LCMP does not necessarily match the sequence number of a corresponding LRQ.
PAGE 431
Example of End-to-End Protocol Packet Exchanges Subsystem Description Data Exchange With Lost Data Figure 17-4 shows a data exchange in which a packet is not received. This problem is usually caused by network congestion and/or line failures and is indicated by a large number of NAKs on the SCF STATS display. Figure 17-4. Lost Data NODE \A NODE \B LRQ (0) LRQ (1) Lost packet LRQ (2) Out-of-sequence (OOS) timeout period NAK (1) LRQ (1) LRQ (2) ACK (3) LCMP (0) LCMP (1) ACK (3) LCMP (2) CDT 013.
PAGE 432
Example of End-to-End Protocol Packet Exchanges Subsystem Description If node \B did not acknowledge node \A’s ENQ, node \A would continue sending ENQs until it reached its Level 4 retry limit or until node \B acknowledged the ENQ, whichever came first. Note. The default OOS timeout is 300 (3 seconds). The OOS timeout can be controlled with the Expand SCF ALTER PATH command or the WAN subsystem SCF ALTER DEVICE command. You can control the Expand subsystem’s retry limit by setting the L4RETRIES modifier.
PAGE 433
Example of End-to-End Protocol Packet Exchanges Subsystem Description Node \B receives the ENQ, responds by resending ACK sequence number 3 to acknowledge the three LRQs, and then sends an LCMP in response to each LRQ. Node \A acknowledges node \B’s LCMPs with ACK sequence number 3. If node \B did not acknowledge node \A’s ENQ, node A would continue to send ENQs until it reached its Level 4 retry limit or until node \B acknowledged the ENQ.
PAGE 434
Layer 4 Send Window Subsystem Description Node \A looks for responses to send to node \B and sends LCMP sequence number 0. This LCMP is a response to a prior request from node \B. When node \B acknowledges the LCMP with ACK sequence number 1, it deallocates its buffer and releases sufficient resources to receive node \A’s LRQ. Node \A resends its initial LRQ (which is now assigned LRQ sequence number 1) along with two more LRQs.
PAGE 435
Routing and Time Factors Subsystem Description Routing and Time Factors This subsection explains how $NCP implements its routing scheme.
PAGE 436
Setting Time Factors Subsystem Description • Expand’s multi-CPU paths are made up of two or more direct paths to the same neighbor that operate in parallel. So the calculation of a multi-CPU path time factor is done in a very similar way as the time factor for a multi-line path (where you have parallel lines). The time factor for a path to a remote (multi-hop) node is calculated as the sum of the time factors for all direct (single-hop) paths that make up the path.
PAGE 437
Negotiating Path Time Factors Subsystem Description RSIZE n has a range of 0 to 186 to designate the line time factor in selecting the best path to other nodes in the network. A smaller number indicates a more desirable path for routing. As always, the actual time factor used for a path between two immediate neighbors is negotiated and the larger of their respective calculations is used.
PAGE 438
Subsystem Description • • Network Routing Table (NRT) and Multiple Path Table (MPT) If two or more routes have the same TF, the route that has the lowest hop count (HC)—the fewest intervening nodes—is selected. Each path between two nodes is one hop. For example, a route that includes one passthrough node has a HC of 2; a route that includes two passthrough nodes has an HC of 3, and so on.
PAGE 439
Network Routing Table (NRT) and Multiple Path Table (MPT) Subsystem Description $NCP sends routing information to the $NCPs at its neighbor nodes at the following times: • • As soon as $NCP becomes aware of a change in the network, such as a line going up or down or a node being added or deleted. During a regular maps exchange. (Maps exchanges are described in Regular Maps Exchanges on page 17-26.
PAGE 440
Calculating Route Time Factors Subsystem Description Regular Maps Exchanges A maps exchange is a periodic sharing of network map information. Maps messages, called distance vector (DV) messages, are exchanged at variable-rate intervals by default. You can specify a fixed five-minute interval exchange by setting the AUTOMATICMAPTIMER modifier. Note. The AUTOMATICMAPTIMER modifier is explained in Section 6, Configuring the Network Control Process.
PAGE 441
Routing Algorithms Subsystem Description Routing Algorithms Routing algorithms determine what and how much routing information $NCP will share with the $NCPs at its neighbor nodes. You can select from two different routing algorithms by setting the ALGORITHM modifier: modified split horizon (MSH) and split horizon (SH). MSH is the default algorithm. Note. ALGORITHM 0 specifies MSH, and ALGORITHM 1 specifies SH. The ALGORITHM modifier is explained in Section 6, Configuring the Network Control Process.
PAGE 442
Routing Algorithms Subsystem Description Figure 17-9. Routing Information With the MSH Algorithm Node \A Node \B TF 4 TF 1 TF 1 TF 1 TF 1 NRT Node \C To Nodes Node \E Node \D \C Via Nodes \B \E \A 2 (2) -- -- \B -- 1 (1) -- \C 1 (1) -- -- \E -- -- 1 (1) CDT 018.
PAGE 443
Routing Algorithms Subsystem Description Split Horizon (SH) When the split horizon (SH) algorithm is used, $NCP tells its neighbor $NCP the bestpath route or the second-best route to a destination node. If the best-path route leads through the neighbor being updated, $NCP will tell its neighbor $NCP its second-best route as long as that route does not lead directly through the neighbor being updated. Figure 17-10 shows the routing information known by node \D when the SH algorithm has been selected.
PAGE 444
Multi-CPU Paths Subsystem Description The disadvantage of the SH algorithm is that it increases the occurrence of loop routing, which results in excessively long routes. Loop routing most often occurs in large, multi-ringed networks. For example, in Figure 17-10, suppose the path fails between node \D and node \E. If a message is sent from node \A to node \E, the Expand subsystem will attempt to reroute traffic in the following sequence: • • Through nodes \B, \A, \C, and \D.
PAGE 445
Multi-CPU Paths Subsystem Description Neighbor Nodes For neighbor nodes, Expand line-handler pairs apply only to each source and destination processor combination, not to entire systems. This method allows traffic between neighbor nodes to be distributed over all the paths in the multi-CPU path. Message order is preserved only between processor pairs instead of between entire systems. $NCP does not establish Expand line-handler pairs with a neighbor node.
PAGE 446
Multi-CPU Paths Subsystem Description Caution. A multi-CPU rebalance can introduce a temporary disruption in the network, similar to but in general less than that caused by an Expand path change. For that reason, it is recommended that rebalances be limited to off-peak hours unless an imbalance is clearly causing immediate problems. The three goals are handled in three separate steps. 1.
PAGE 447
Multi-CPU Routing Examples Subsystem Description • • • • When a new path comes up. (This is similar to what happens in normal paths when a new path that has a lower TF is discovered.) At configurable times during the day. You can use the SCF ALTER PROCESS, AUTOREBALANCE command to specify when rebalancing should occur. Both the time of day and the interval between rebalance attempts can be specified, allowing you to schedule a rebalance when traffic is minimal. Immediately.
PAGE 448
Multi-CPU Routing Examples Subsystem Description Figure 17-11. Network Containing Normal Paths and Multi-CPU Paths Node \B CPU 1 LH Node \A LH PRCB PRCA CPU 0 CPU 2 CPU 0 CPU 1 CPU 2 CPU 3 LH CPU 4 LH LH LH Multi-CPU Path 1 Node \C CPU 0 LH CPU 1 LH PRCC Node \D CPU 2 CPU 1 LH LH CPU 0 LH CPU 3 LH LH CPU 2 Multi-CPU Path 2 LH Node \F Node \E LH CPU 0 LH CPU 1 PRCF LH CPU 0 CPU 1 CPU 2 CPU 2 LH CPU 3 CDT 053.
PAGE 449
Subsystem Description Multi-CPU Routing Examples Combination 1: Local Source Node and Neighbor Destination Node In this scenario, the source node is the local node and the destination node is a neighbor; a message is sent directly from one node to the other. When the first message destined for each processor in the neighbor node is sent, the originating processor selects a local path to the destination node and selects a pair of Expand line-handlers for the source and destination processor combination.
PAGE 450
Subsystem Description Multi-CPU Routing Examples RPT entry in all processors in the system. If a message is received from a neighbor node and no RPT entry exists, the message is dropped. For example, in Figure 17-11, when $NCP on node \A first detects the existence of node \C, $NCP sends a Connect Request message to node \B which is forwarded through multi-CPU path 1 to node \C.
PAGE 451
Message Handling and Buffer Allocation Subsystem Description Message Handling and Buffer Allocation This subsection presents a high-level overview of how data is sent and received over an Expand network and how incoming and outgoing data is buffered. It is necessary to understand this information to effectively configure, manage, and troubleshoot an Expand network. This subsection describes the following topics: • • Outgoing Traffic Flow on page 17-37 Incoming Traffic Flow on page 17-41 Note.
PAGE 452
Outgoing Traffic Flow Subsystem Description Locally Originated Traffic Flow Figure 17-12 illustrates the path of a locally originated outgoing message. Figure 17-12.
PAGE 453
Outgoing Traffic Flow Subsystem Description password exists for the destination node, the request is completed with an error (filesystem error 48, security violation) and is not sent. If the COMPRESS_ON modifier is set, the Expand line-handler process attempts to compress the data in the message. When compression is configured, groups of consecutive zeros (0), spaces, and NULLs are replaced with indicator and count values.
PAGE 454
Outgoing Traffic Flow Subsystem Description Note. Requests are formatted into request data packets, or LRQs. Replies are formatted into reply data packets, or LCMPs. LRQs and LCMPs are explained in Protocol Packet Types on page 17-12. $NCP and Passthrough Traffic Flow Figure 17-13 illustrates the path of outgoing $NCP and passthrough traffic. Figure 17-13.
PAGE 455
Incoming Traffic Flow Subsystem Description When passthrough and $NCP traffic is queued to the outgoing list, it occupies buffer space in the Expand line-handler process buffer pool. $NCP formats $NCP messages into packets before sending them to the appropriate Expand line-handler process for transmission. Passthrough traffic is already in the form of packets; it is not reassembled into messages before being forwarded to the destination node. Note.
PAGE 456
Incoming Traffic Flow Subsystem Description Figure 17-14.
PAGE 457
Incoming Traffic Flow Subsystem Description Request Packets An incoming request packet, or an LRQ, is a fragment of a request message destined for a process at the local node. The first LRQ includes the length of the total message, in bytes. The Expand line-handler process reserves memory from its buffer pool for the total message based on the length information contained in the first packet. Note. LRQs are also described in Protocol Packet Types on page 17-12.
PAGE 458
Incoming Traffic Flow Subsystem Description Once the reply message is successfully processed, the message system routes the reply message to the appropriate process, and the Expand line-handler process releases the buffer pool used by the reply message. $NCP and Passthrough Packets An incoming $NCP packet is a packet received from the $NCP at a neighbor node and destined for the $NCP of the local node.
PAGE 459
Message Buffering Subsystem Description Message Buffering The previous subsection showed that Expand line-handler processes buffer incoming and outgoing requests so that data can be transferred between processes on different nodes. This subsection describes in greater detail the data space allocated to the Expand line-handler process for message transfer and how you can affect the size of that buffer space.
PAGE 460
Global Variables Subsystem Description Global Variables The global variables space contains the Expand subsystem software global variables. The Expand subsystem determines how much global variables space to allocate according to the number of lines in a path controlled by the Expand line-handler process. Stack The Expand subsystem allocates 700 words to the stack. Control Blocks The Expand subsystem preallocates space for many data structures that are likely to be used during normal operation.
PAGE 461
Shared Memory Area for QIO Subsystem Description The SCF attribute EXTMEMSIZE n allows you to specify the base size of extended memory for the pool, from a default of 2 megabytes to as much as 32 megabytes. This extra memory will be of invaluable help to applications such as the Remote Database Facility (RDF) which in the past suffered from memory pool problems and thus reduced performance.
PAGE 462
Expand-to-NAM Interface Subsystem Description Expand-to-NAM Interface This subsection describes how Expand-over-NAM and Expand-over-ServerNet linehandler processes access a network access method (NAM) interface. The information presented in this subsection will help you effectively configure, manage, and troubleshoot an Expand network that includes X.25, SNA, or ServerNet connections.
PAGE 463
Connection Establishment Subsystem Description Connection Establishment Figure 17-16 illustrates the events that occur when Expand-over-NAM and Expandover-ServerNet line-handler processes successfully establish a connection through a NAM interface. Figure 17-16.
PAGE 464
Connection Establishment Subsystem Description The Expand-over-NAM or Expand-over-ServerNet line-handler process accesses a subdevice by sending a bind request to the NAM process. A bind request is roughly equivalent to an OPEN procedure.
PAGE 465
Sending and Receiving Data Subsystem Description connect request. If you specify the MAXRECCONNECTS modifier, you can also control what happens after the reconnect limit has been reached by specifying the AFTERMAXRETRIES_PASSIVE or AFTERMAXRETRIES_DOWN modifier. Passive Connect Request When the Expand-over-NAM or Expand-over-ServerNet line-handler process issues a passive connect request, the NAM waits for an incoming connect request.
PAGE 466
Expand-to-IP Interface Subsystem Description Expand-to-IP Interface This subsection describes how the Expand-over-IP line-handler process accesses an Internet Protocol (IP) network. You should be familiar with the information presented in this subsection before attempting to configure, manage, or troubleshoot an Expand network that includes IP connections.
PAGE 467
Expand-over-IP Connection Establishment Subsystem Description Each NonStop TCP/IP process appears to an IP network as a separate host and is associated with a separate IP address. An IP address is a 4-octet (32-bit) numeric value identifying a particular network (network address portion) and a local host on that network (local address portion). A NonStop TCP/IP process can be associated with more than one IP address.
PAGE 468
Expand-over-IP Connection Establishment Subsystem Description When the remote Expand-over-IP line-handler process receives the Connect Command frame, it responds with a Connect Response frame. When the response is received, the local Expand-over-IP line-handler process considers the line to be up. Various path parameters are then exchanged with the remote Expand-over-IP linehandler process.
PAGE 469
Sending and Receiving Data Subsystem Description limit has been reached by specifying the AFTERMAXRETRIES_PASSIVE or AFTERMAXRETRIES_DOWN modifier. Sending and Receiving Data Once a connection has been established, the local and remote Expand-over-IP linehandler processes communicate through their associated NonStop TCP/IP or TCP6SAM processes using the QIO mechanism.
PAGE 470
Forwarding Expand-over-IP Packets to Other Expand Line-Handler Processes Subsystem Description Forwarding Expand-over-IP Packets to Other Expand LineHandler Processes Packets received by an Expand-over-IP line-handler process can be forwarded to another type of Expand line-handler process, either on the same processor or on a different processor.
PAGE 471
Subsystem Description Expand-to-ATM Interface Expand-to-ATM Interface This subsection describes how the Expand-over-ATM line-handler process accesses an Asynchronous Transfer Mode (ATM) network. You should be familiar with the information presented in this subsection before attempting to configure, manage, or troubleshoot an Expand network that includes ATM connections.
PAGE 472
Subsystem Description Expand-over-ATM Connection Establishment SVC Connections An SVC is a dynamically established virtual circuit. Each SVC is automatically assigned an SVC name by the ATM3SA when the circuit is established. An SVC is described by the SVC object, which is subordinate to the LINE object.
PAGE 473
Subsystem Description Expand-over-ATM Connection Establishment The default connect method is active connect. You can cause the Expand-over-ATM line-handler process to use the active connect method by specifying the CONNECTTYPE_ACTIVEANDPASSIVE modifier. Active Connect Request When the Expand-over-ATM line-handler process issues an active connect request, it attempts to initiate a connection by sending a Connect Command frame to the remote Expand-over-ATM line-handler process.
PAGE 474
Subsystem Description Sending and Receiving Data attribute does not correspond to an Expand modifier and can therefore be changed only by using the SCF interface to the Expand subsystem.) You can limit the number of times the Expand-over-ATM line-handler process will send a Connect Command frame by specifying the MAXRECONNECTS modifier.
PAGE 475
Subsystem Description Forwarding Expand-over-ATM Packets to Other Expand Line-Handler Processes Forwarding Expand-over-ATM Packets to Other Expand LineHandler Processes Packets received by an Expand-over-ATM line-handler process can be forwarded to another type of Expand line-handler process, either on the same processor or on a different processor. Packet forwarding is performed via the message system; this allows servers without Expand-over-ATM line-handler processes to access an ATM network.
PAGE 476
Subsystem Description Multipacket Frame Feature Multipacket Frame Feature The multipacket frame feature is a performance enhancement designed to increase throughput and processor efficiency on all connection types. This subsection briefly describes how the multipacket frame feature works so that you can effectively configure and use this feature in your network.
PAGE 477
Subsystem Description • • • Constructing Multipacket Frames If the Layer 2 protocol is a NAM interface, each Expand packet is treated as a separate NAM message. If the Layer 2 protocol is NETIP, each Expand packet is treated as a separate UDP frame. If the Layer 2 protocol is NETATM, each Expand packet is treated as a separate ATM frame. Figure 17-19 shows how Expand packets are sent over a direct-connect (HDLC) connection when the multipacket frame feature is not selected.
PAGE 478
Subsystem Description Path Initialization In Figure 17-20, a message is passed to the direct-connect line-handler process with the multipacket frame feature selected. The direct-connect line-handler process still fragments the message into six Expand packets but now constructs one large multipacket frame to hold all six packets. If the entire multipacket frame fits inside one HDLC-type frame, it is sent across the line in one frame.
PAGE 479
Subsystem Description Multipacket Frame Configuration Multipacket Frame Configuration The FRAMESIZE modifier determines the maximum Expand packet size, in bytes, according to the following formula: packet_size = ( FRAMESIZE - 4 ) * 2 For example, the default value for the FRAMESIZE modifier is 132, establishing a maximum packet size of 128 words (or 256 bytes). The FRAMESIZE modifier must be the same value for every Expand line-handler process in the network. Note.
PAGE 480
Subsystem Description Variable Packet Size Feature Variable Packet Size Feature The variable packet size feature is a performance enhancement designed to improve bulk transfers over all connection types. The variable packet size feature effectively overrides the packet size calculated from the FRAMESIZE modifier value by allowing you to configure a maximum packet size, which is used for both single-packet and multipacket frames, on a per-path basis.
PAGE 481
Subsystem Description • • • Mixing Extended and Nonextended Packets The variable packet size feature does not provide any benefit on paths configured with the L4EXTPACKETS_OFF modifier, which specifies that the extended 64-byte packet header format not be used. Nonextended frames are not fragmentable and therefore must use the network-wide FRAMESIZE modifier value.
PAGE 482
Subsystem Description Considerations for Paths Using the Variable Packet Size Feature and the Multipacket Frame Feature Considerations for Paths Using the Variable Packet Size Feature and the Multipacket Frame Feature The main difference between the variable packet size feature and the multipacket frame feature is that the multipacket frame feature benefits users who send many small concurrent requests, while the variable packet size feature benefits users who send large blocks of data (bulk transfers).
PAGE 483
Subsystem Description Congestion Control Feature Congestion Control Feature Congestion in a network occurs when performance on a connection degrades due to the saturation of a resource that is needed to deliver data from the source to the destination. Congestion control mechanisms regulate system resources in order to avoid network bottleneck and resource contention situations.
PAGE 484
Subsystem Description Congestion Control Feature Figure 17-22. Congestion Control Not Enabled Node \B Node \A Node \C On Off Off On L4CONGCTRL_ON L4CONGCTRL_OFF L4CONGCTRL_ON Congestion Control On Congestion Control On CDT 014.CDD Nodes that support congestion control are compatible with nodes that do not. However, connections between such nodes will not use congestion control. In Figure 17-23, congestion control is enabled on nodes \A and \C but is not enabled on node \B.
PAGE 485
Subsystem Description Congestion Control Configuration The congestion control feature uses end-to-end mechanisms for congestion control and error-recovery. It does not provide any mechanisms for indicating congestion along intermediate nodes. Congestion Control Configuration You select the congestion control feature by specifying the L4CONGCTRL_ON modifier. This modifier enables the congestion control mechanism for sending packets on a specific path.
PAGE 486
Subsystem Description Multi-CPU Feature Multi-CPU Feature The Expand multi-CPU feature enables you to spread the communications load over multiple processors by connecting multiple Expand line-handler process, each in a separate processor, between two adjacent nodes.
PAGE 487
Subsystem Description Multi-CPU Considerations modifier and one or more the existing paths are in a multi-CPU path, then the new path joins the multi-CPU path. If there is no preexisting multi-CPU path, then a multi-CPU path is created with the new path as its sole member. When a path comes up, it negotiates its multi-CPU membership with the Expand linehandler process on the other end of the connection.
PAGE 488
Subsystem Description Multi-CPU Considerations Expand Configuration and Management Manual—529522-002 17-74
PAGE 489
Part V.
PAGE 490
Part V.
PAGE 491
18 Managing the Network This section explains how to access network resources, set up network security, and monitor, reconfigure, and control an Expand network.
PAGE 492
Managing the Network Using TACL to Manage Remote Files Using TACL to Manage Remote Files One of the major features of the Expand subsystem is network transparency. Because access to the network is transparent to the user, the Expand subsystem does not include its own network commands. This subsection describes how to use TACL commands to manage remote files. Note. Selected TACL commands are described in this subsection.
PAGE 493
Managing the Network Changing Your Default Values Changing Your Default Values Each user on the system has two sets of default values: current default values and saved default values. Saved default values are in effect when you log on. Current default values define your present location or frame of reference in the system and network. You can move around on the system and network by changing the current system, volume, and subvolume defaults.
PAGE 494
Managing the Network Gaining Access to Remote Nodes Note. Changing the current default node does not log you onto the other node. To log onto a node other than the one where your current TACL process is running, you must first start a remote TACL process on that node. Logging on to a remote node is described in Starting and Quitting a Remote TACL Process on page 18-4.
PAGE 495
Managing the Network Gaining Access to Remote Nodes is part of a network that includes the \HERST node, you can start a TACL process on \HERST by entering the following command: \HERST.TACL The TACL program returns the initial TACL prompt, and you can now log onto the \HERST system. A remote TACL started this way does not have a backup process. If you want the remote TACL process to run as a process pair, enter the following command instead of the previous command: \HERST.
PAGE 496
Managing the Network Setting Up Network Security For example, to run a program named MYPROG on the remote node \CITY using an explicit RUN command, you would type the following command: RUN \CITY.MYPROG To run the same program using an implicit RUN command, you would type the following command: \CITY.MYPROG When you run a program on a remote node, the default volume and subvolume names remain in effect. Unless you use the SYSTEM command to change the default node, the local node remains the default.
PAGE 497
Managing the Network Remote File Security Remote File Security A user on node \WEST who wants to access a file (including a disk file, device, or process) on a node \EAST must satisfy the following requirements: • • • The user must also be established as a user on node \EAST. The user must have matching remote passwords established on both nodes. To access a disk file, the user on node \WEST must have authority to access the file on node \EAST as a remote accessor.
PAGE 498
Managing the Network Establishing Remote Passwords The allow-access password for ADMIN.BILL for \WEST from all other nodes is SHAZAM. At node \EAST, the following commands are entered: logon admin.bill remotepassword \west, shazam The user at node \EAST entered the matching password and now has remote access to node \WEST as ADMIN.BILL. ADMIN.BILL, logged on at node \EAST, does not have the same status at \WEST as does the ADMIN.BILL at \WEST. Because ADMIN.
PAGE 499
Managing the Network Remote Process Security The following command removes all of the user’s remote passwords: remotepassword • • Request-access passwords and allow-access passwords can be specified at any time. Remote access is permitted as soon as both remote passwords are defined (provided they match). Remote passwords are independent of local passwords. In the preceding example, ADMIN.BILL could prevent unauthorized persons from logging on as ADMIN.
PAGE 500
Managing the Network Global Remote Passwords secured “OOOO” (local owner only) along with other files that are only accessible locally. A remote user can be prevented from becoming a local user if the local super ID specifies “A” (any local user) as the execute security for the TACL program file. This prevents anyone on a remote node from starting a TACL process on the local node.
PAGE 501
Managing the Network Remote Super ID User like the one used in the previous example allows certain users to log on as NET.WEST. Subnetworks implemented in this manner can overlap or include one another. \CHICAGO might be accessible from \NEWYORK by logging on as NET.EAST, and from \PHOENIX by logging on as NET.MIDWEST. Similarly, each system in the network might have a user called NET.GLOBAL, who is allowed to access every other node.
PAGE 502
Managing the Network • • Monitoring Network Activity Auditing of file access, logon/logoff, and changes to security or security controls Controlled file and process creation Safeguard is described in the Safeguard Administrator’s Manual. Monitoring Network Activity Network monitoring includes gathering statistical information, checking the status of hardware and software components, and displaying configuration values.
PAGE 503
Managing the Network Displaying $NCP Information Table 18-1. Expand SCF Commands for $NCP Information (page 2 of 2) SCF Command Information Reported INFO PROCESS $NCP, PATHSETS Displays the NCP pathmap information similar to the LINESET command, but displays it in a different format. This format displays both the line-handler LDEV and name, as well as the other information already in the LINESET command.
PAGE 504
Managing the Network Displaying Expand Line-Handler Process Information Table 18-2. WAN SCF Commands for $NCP Information SCF Command Information Reported INFO DEVICE $ZZWAN.#NCP Displays the primary and backup processors, type, record size, object file, and profile used by $NCP. The DETAIL option can be used to display device-specific modifiers and modifier values. INFO PROFILE $ZZWAN.#ncp_profile Displays a list of the modifiers and modifier values contained in the profile used by $NCP.
PAGE 505
Managing the Network Displaying Expand Line-Handler Process Information Table 18-4 lists the subtype values associated with single-line Expand line-handler processes. Table 18-4. Subtype Values for Single-Line Line-Handler Processes Line Type Subtype Direct-connect 5 Satellite-connect 5 Expand-over-NAM 0 Expand-over-IP 0 Expand-over-ATM 0 Expand-over-ServerNet 4 Table 18-5 lists the subtype values associated with multi-line paths (path and line logical devices). Table 18-5.
PAGE 506
Managing the Network Displaying Expand Line-Handler Process Information Table 18-6. WAN SCF Commands for Expand Line-Handler Process Information (page 2 of 2) SCF Command Information Reported STATUS DEVICE $ZZWAN.#device_name Displays the dynamic state, logical device (LDEV) number, and primary and backup process identification numbers (PINs) for a selected Expand linehandler process.
PAGE 507
Managing the Network Displaying Expand Line-Handler Process Information Table 18-7. Expand SCF Commands for Line Information (page 2 of 2) SCF Command Information Reported STATUS LINE $device_name Displays status information for a selected Expand linehandler process. Information displayed includes the summary state of the line, primary process ID (PID), and backup process ID (PID).
PAGE 508
Managing the Network Starting and Stopping Tracing Starting and Stopping Tracing The Expand subsystem SCF TRACE command allows you to select the records that you want written to a disk file. You can then use PTrace commands to select records to be formatted and sent to an output device. Table 18-9 lists the Expand subsystem SCF commands that can be used to start and stop tracing. Table 18-9.
PAGE 509
Managing the Network Reconfiguring the Network Reconfiguring the Network Network reconfiguration tasks include the following: • • • • • • • • Adding and Deleting Expand Line-Handler Processes Adding and Deleting $NCP Changing $NCP Modifiers Changing Expand Line-Handler Process Modifiers Changing Profiles Adding Nodes to the Network Removing Nodes From the Network Changing System Names and Numbers Note.
PAGE 510
Managing the Network Changing Expand Line-Handler Process Modifiers Changing Expand Line-Handler Process Modifiers You can use the WAN subsystem SCF ALTER DEVICE command to change any modifier or modifier value in the device record for a specific Expand line-handler process. You can use the Expand subsystem SCF ALTER LINE and ALTER PATH commands to change certain Expand line-handler process modifiers and modifier values.
PAGE 511
Managing the Network Adding Nodes to the Network Creating and starting Expand line-handler processes is explained in detail in Section II, Configuring the Expand Subsystem. Note. Before you can start an Expand line-handler process, other processes might need to be present and running in your system. Refer to Section II, Configuring the Expand Subsystem for more information.
PAGE 512
Managing the Network Removing Nodes From the Network Removing Nodes From the Network This subsection explains how to remove a node from the network using the management commands described in the preceding subsections. This explanation is presented in two steps.
PAGE 513
Managing the Network Changing System Names and Numbers Step 1: Save the current configuration file As a precaution, use the SCF SAVE command to save the current configuration files on the duplicate nodes. For example, the following command saves the configuration file at $SYSTEM.ZSYSCONF.CONF0101: -> SAVE CONFIG 1.1 The SCF SAVE command is described in detail in the SCF Reference Manual for GSeries RVUs.
PAGE 514
Managing the Network Changing System Names and Numbers Step 4: Change the system name and/or system number You must change the system name and/or system number of one of the duplicate nodes. To change the system name or number, use the Kernel subsystem SCF ALTER command.
PAGE 515
Managing the Network Controlling the Network Step 6: Perform a system load Because the attributes that change the system name and number are stored in a SEEPROM in the Integrity NonStop NS-series server backplane, changes to them will not take effect until you perform a system load. Note. You must perform the system load using the Start System button or the Start System command (under the Operations menu) in OSM. For more information, see the online help within OSM.
PAGE 516
Managing the Network Stopping and Starting Lines and Paths Stopping and Starting Lines and Paths The SCF interface to the Expand subsystem provides commands to control lines and paths. Table 18-11 describes each of these commands and the actions they perform. Table 18-11. Expand SCF Control Commands SCF Command Action Performed ABORT LINE $device_name Terminates the operation of a line as quickly as possible. Only enough processing is done to ensure the security of the subsystem.
PAGE 517
Managing the Network Rebalancing Multi-CPU Paths Table 18-12. Expand SCF Commands for Switching Processors SCF Command Action Performed PRIMARY PROCESS $device_name Causes the backup process to become the primary process, or the primary process to become the backup process, for a selected Expand line-handler process. PRIMARY PROCESS $NCP Causes the backup process to become the primary process, or the primary process to the backup process, for $NCP.
PAGE 518
Managing the Network Rebalancing Multi-CPU Paths Expand Configuration and Management Manual—529522-002 18-28
PAGE 519
19 Tuning This section provides guidelines for improving network performance and describes the tools available for measuring performance.
PAGE 520
Tuning Performance Factors Performance Factors This subsection describes the factors that can be adjusted to improve Expand linehandler process performance and processor utilization. Performance factors, and their relative effect on the tuning goals described earlier in this section, are shown in Table 19-1. How to Use the Performance Factors Table To use Table 19-1, vertically scan the Tuning Goals columns.
PAGE 521
Tuning Multipacket Frame Size Multipacket Frame Size The multipacket frame feature is designed to reduce processor use at nodes where the workload is high and the configured frame size must remain unchanged. This feature enables multiple packets to be placed in a single frame (instead of a single packet in a single frame). The multipacket frame feature is supported for all line types.
PAGE 522
Tuning Multipacket Frame Size Figure 19-1. Throughput With and Without Multipacket Frames Kilobits per second 1800 1600 1400 1200 1000 800 600 400 200 0 132 516 744 Framesize (in words) Single-packet Multipacket VST074.vsd Processor Use and Message Size Multipacket frames can improve the processor efficiency of all line types.
PAGE 523
Tuning Variable Packet Size Multipacket Frame Configuration The multipacket frame size is determined by the value assigned to the PATHBLOCKBYTES modifier. When the variable packet-size feature (PATHPACKETBYTES modifier) is used, the Expand subsystem should be able to send a full variable-size packet inside a multipacket frame. For this reason, the PATHBLOCKBYTES modifier must be set to a value greater than or equal to the PATHPACKETBYTES modifier value.
PAGE 524
Tuning Variable Packet Size transfers are much more expensive to form into small packets and route in multihop networks. Extended Packet Format The extended packet format (L4EXTPACKETS_ON modifier) provides a means to fragment packets in transit across the network. The extended packet format must be enabled for the variable packet-size feature to function.
PAGE 525
Tuning Application Message Size Note. If you use the Expand subsystem SCF ALTER LINE command to set the L2TIMEOUT modifier, you must convert the result of this formula to a time interval. For example, if the result is 300 (3 seconds), you will enter the following command: ALTER LINE $device_name, L2TIMEOUT 3.00 For more information about configuring the variable packet-size feature, refer to Variable Packet Size Feature on page 17-66.
PAGE 526
Tuning Application Message Size As shown in Figure 19-2, QIO recognizes that the data is to be sent to an application that is not on the local node, and it routes the request to the appropriate Expand linehandler process. If the Expand packet size is large enough to hold all of the message from the application, the Expand line-handler process puts the message into a single packet.
PAGE 527
Tuning Packet Format Packet Format The Expand subsystem enforces a minimum value of 1024 bytes for the variablepacket size (set with the PATHPACKETBYTES modifier). The default value for PATHPACKETBYTES (1024 bytes) yields the same data-per-packet percentage as nonextended packets with a frame size of 132 words. Table 19-2 compares the data-per-packet percentages for nonextended packet (L4EXTPACKETS_OFF modifier) and extended packet (L4EXTPACKETS_ON modifier) header formats.
PAGE 528
Tuning Layer 2 Window Size Expand-Over-IP Connections Expand-over-IP line-handler processes use the User Datagram Protocol (UDP) services provided by a NonStop TCP/IP process to transmit data across an Internet Protocol (IP) network. Because data transfer with UDP is not guaranteed, the Expand End-to-End protocol is used to achieve reliable communications for Expand-over-IP connections.
PAGE 529
Tuning Processor Type Processor Type The processing power of the Integrity NonStop NS-series servers (through which a message is transmitted) determines throughput if there are no bottlenecks in the other components of the network. The relative processor power factors are a good starting point for estimating Expand line-handler process performance limits. Process location and load balancing within a system can also have a major impact on network performance.
PAGE 530
Tuning NAM Interface NAM Interface When a NAM interface is used, Layer 2 functions are managed by the NAM process, thus reducing the load on the Expand line-handler process. Although the Expand linehandler process has a potentially greater upper throughput limit when it uses a NAM interface, overall system processor requirements are not reduced because some of the workload is shifted to the NAM process.
PAGE 531
Tuning Multi-Line Paths Multi-Line Paths The multi-line path feature enables you to configure eight parallel lines between the same two nodes. The advantages of multi-line paths include increased fault-tolerance and additional bandwidth. The main disadvantage of multi-line paths is increased processor overhead, which occurs primarily because extra processing must be done to select the best line for each frame transmitted and to guarantee sequencing of packets received across multiple lines.
PAGE 532
Tuning Multi-CPU Paths Window Size If a line is added to a path using a different protocol or telecommunications network at Layer 1, the path might have different delay characteristics from the original single line. It might be necessary to change the Layer 2 window size (TXWINDOW modifier) to minimize any delays introduced by switches or new protocols. In some situations, the HDLC Extended protocol might be advantageous on terrestrial links.
PAGE 533
Tuning Multi-CPU Paths The main disadvantage of the Expand multi-CPU feature is that its advantages are available only when traffic fits a certain pattern. For example, if most traffic occurs between the same two nodes—or if these nodes are direct neighbors and traffic is sent between the same two processors in one direction—then the Expand multi-CPU feature cannot spread the load effectively.
PAGE 534
Tuning Multi-CPU Paths Load Balancing When a multi-CPU path initially assigns paths to each pair of endpoints, the traffic pattern is usually not yet known. Load balancing is used to correct this problem as more information is gathered by moving Expand line-handler process pairs from more heavily loaded paths to more lightly loaded paths within the multi-CPU path. A slight disruption occurs in message transfer occurs when pairs are changed.
PAGE 535
Tuning Multi-CPU Paths Superpath Load Distribution The Superpaths feature does not distribute the load equally over all paths. A superpath distributes the load based on three criteria: • • • CPU Matching Load Factor Balancing Pair Count Balancing CPU Matching This takes effect when the two systems are directly connected with a superpath; that is, they are direct neighbors.
PAGE 536
Tuning Multi-CPU Paths Load Factor Balancing If there are no matching CPUs, then the load would be distributed based on the load factor of the paths in the superpath. If a process on \A in CPU 0 is communicating with a process on \B in CPU 2, the line handler chosen is based on the load factor of the two lines. Once the CPU pair has been established, that line handler is used for all communication between the two CPUs.
PAGE 537
Tuning Multi-CPU Paths In reference to \A, \B is a neighbor and \C and \D are non-neighbors. When \A makes a connection to \C, the load is not distributed over different paths, but only one path is used for all traffic to \C. In order to make sure that the correct path is used, \B records which line handler is used for \A's connection to \C. This is done by setting an entry in the reverse pairing table so \B knows which line handler to send the packets from \C to \A.
PAGE 538
Tuning Network Topology load factor to each other line handler and predict the resulting change in load factors. Choose the single move that results in the lowest predicted load factor spread, put it on the output change list, update the load factors according to the predicted changes, and check the new load factor spread value. This is continued until the load factor spread is less than 0.5 or no moves can be found that improve the load factor spread. 3. Lastly, the pair counts are balanced.
PAGE 539
Tuning Summary of Tuning Strategies Passthrough data has a 4-to-1 priority over locally originated data. This ratio is tuned fairly well for small passthrough packets. If all nodes in a route are configured for a large variable packet size (PATHPACKETBYTES modifier) such as 4095 bytes, the intermediate nodes can send up to 16 Kbytes of passthrough traffic between packets of a locally originated message.
PAGE 540
Tuning Measuring and Mapping an Expand Network Measuring and Mapping an Expand Network Effective network tuning must begin with an accurate picture of the network and its significant traffic. All networks are fully defined topologically by their nodes and links. Network traffic is fully defined by the intensity of the traffic, the service times of the working components, and the capacities of passive resources such as buffers and queues. These parameters can all be measured by HP utilities.
PAGE 541
Tuning Using Measure Using Measure The Expand components reported on by Measure are called entities. Once you have mapped your network and selected the nodes you want to measure, you can characterize the network traffic by measuring the following entities: • • • • • SYSTEM Entity NETLINE Entity LINE Entity PROCESS Entity CPU Entity Note. The values shown in the following diagrams are in units per second (the Measure SET REPORT RATE ON command was used).
PAGE 542
Tuning Using Measure NETLINE Entity The NETLINE entity reports several values that you need when modeling traffic on a path. NETLINE reports part of the path performance associated with a particular Expand line-handler process, logical device, and ServerNet wide area network (SWAN) concentrator. The number of data bytes received is shown in Din4-Bytes and the number of bytes sent is shown in Dout4-Bytes. NETLINE also shows the distribution of message sizes in message-size ranges.
PAGE 543
Tuning Using Measure Example 19-3. LINE Entity Display Comm Line $X25TAH Device Type 61 Subdevice Type 63 Logical Device 170 Trackid SWAN24 Clip 2 Line 1 Local System \COWBOY From 7 Feb 1997, 10:23:51 For 44.3 Seconds Requests Write-busy-Time Read-busy-Time Input-Bytes Input-Data-Bytes Transactions 6.03 1.21 % 85.52 % 946.94 770.93 Retries Writes Reads Output-Bytes Output-Data-Bytes Response-Time 11.97 12.
PAGE 544
Tuning Using Measure Example 19-4. PROCESS Entity Display Process 3,14 ($B30S) Pri 199 Pg Size 16384 Bytes Program $SYSTEM.SYS01.LHOBJ (Native) Userid 255,255 Creatorid 255,255 Ancestor 1,275 ($ZZWAN) Local System \TAHITI From 7 Feb 1997, 11:53:36 For 3.3 Minutes Cpu-Busy-Time Mem-Qtime Page-Faults Pres-Pages-Qtime Ext-Segs-Qtime Recv-Qtime Messages-Sent Sent-Bytes Returned-Bytes MQC-Allocations MQCs-Inuse-Qtime Checkpoints File-Open-Calls UCL-Qtime Accel-Busy-Time TNSR-Busy-Time Begin-Trans 31.
PAGE 545
Tuning Using Measure Determining Processor Use To determine the processor use of an Expand line-handler process, you must add a fraction of the Intr-Busy-Time count from the processor where the Expand line-handler process is running to the Cpu-Busy-Time value for the Expand line-handler process.
PAGE 546
Tuning Measuring Passthrough Traffic Measuring Passthrough Traffic Although passthrough traffic is reported in the SYSTEM entity (SENT-FWD and RCVDFWD counters), Measure does not directly account for the source and destination of passthrough traffic when examining a path. An Expand line-handler process only sees the node to which it is connected. The only way to map the passthrough traffic accurately is to know the topology of your network and to measure each Expand linehandler process on each node.
PAGE 547
Tuning Example 1: Changing Packet Size Example 19-6. SCF PATH STATS Display -------------------- LEVEL 4 MESSAGE HISTOGRAM --------------------------<= 64 .. 71027 <= 128 .. 25609 <= 256.. 4211 <= 512 .. 1676 <= 1024 .. 1466 <= 2048.. 370 <= 4096 .. 179 > 4096 .. 2288 -------------------- LEVEL 4 / LEVEL 3------------Average--------Average--Packets Forwards Links Packets/Block Bytes/Block Sent 230144 0 24888 1.0 238 Rcvd 94921 0 28559 1.0 178 L4 Packets Discarded.........
PAGE 548
Tuning Example 1: Changing Packet Size per-message processor cost is a constant no matter what packet size is used since the number of messages and the size of the messages is the same in both cases. For messages sent, changing the packet size to 1024 bytes improves the packets-perlink ratio by about five times.
PAGE 549
Tuning Example 2: Reducing Passthrough Traffic Figure 19-6. Packet Size/Bandwidth Comparison Bits per message 20000 15000 10000 5000 0 256 1024 2048 Packet Size (Bytes) VST029.vsd Example 2: Reducing Passthrough Traffic It is common for the role of a node in an Expand network to change over a period of time.
PAGE 550
Tuning Example 2: Reducing Passthrough Traffic Notice that the source and destination of passthrough traffic cannot be identified from Measure data. Example 19-7.
PAGE 551
Tuning Example 2: Reducing Passthrough Traffic Example 19-8. Passthrough Traffic in a Network Local System \HERE Frames % Send % Send Rcv % Rcv Remote Msgs Rcvd Frames Frames Total Pass Pass Pass Pass System Sent Sent Rcvd Rcvd Hops Thru Thru Thru Thru ---------------------------------------------------------------------------\OAHU 40856 87030 81732 9.87% 4 522180 10.14% 490392 14.37% \CANTON 39259 79444 78637 9.50% 4 476664 9.25% 471822 13.83% \SAIPAN 10438 76652 77156 9.32% 1 0 0.00% 0 0.
PAGE 552
Tuning Example 2: Reducing Passthrough Traffic Expand Configuration and Management Manual—529522-002 19-34
PAGE 553
20 Troubleshooting To quickly and efficiently identify and resolve network problems, HP recommends that you use a standard network troubleshooting methodology.
PAGE 554
Troubleshooting SCF SCF The Subsystem Control Facility (SCF) interface to the Expand subsystem provides several commands to help you determine the normal operation of Expand line-handler processes. The SCF STATS command displays Layer 4 and Layer 2 statistical information. The SCF STATUS command displays information about the status of an object, such as its state (STOPPED, STARTING, or STARTED).
PAGE 555
Troubleshooting Identifying Network Problems following manuals: ASAP Client Manual, ASAP Server Manual, ASAP Extension Manual, and ASAP Migration Guide for NSX and OMF Users. Identifying Network Problems There are a number of sources from which to obtain information to identify a network problem. Many of these sources are the same as those used to verify normal system operation.
PAGE 556
Troubleshooting User Complaints User Complaints Most troubleshooting starts with user complaints, which can result from either application or hardware problems. The best approach is always to check the obvious first.
PAGE 557
Troubleshooting SCF Commands Example 20-1. SCF LISTDEV Display LDev Name PPID BPID Type 26 66 68 71 87 88 89 124 2,13 2,14 1,18 2,11 2,10 2,13 2,13 1,23 3,13 3,12 0,17 3,15 3,16 3,13 3,13 0,23 (63,1 ) (63,6 ) (63,0 ) (63,0 ) (63,5 ) (63,6 ) (63,6 ) (63,30) $PBAL4 $A10 $IPCOW $EX25COW $B30S $B21 $B20 $ZEXP RSize Pri Program 0 12 3 12 12 12 12 132 199 199 199 199 199 199 199 180 \TAHITI.$SYSTEM.T9057.LHOBJ \TAHITI.$SYSTEM.SYS01.LHOBJ \TAHITI.$SYSTEM.SYS01.LHOBJ2 \TAHITI.$SYSTEM.SYS01.
PAGE 558
Troubleshooting SCF Commands Example 20-2. SCF STATS Display 3-> stats line $b30s EXPAND Stats LINE $B30S, PPID ( 2, 10), BPID ( 3, 16) Resettime... FEB 18,1997 10:38:12 Sampletime...
PAGE 559
Troubleshooting SCF Commands The SCF INFO PROCESS $NCP LINESET command is useful for displaying the status of a selected path and the lines in that path. The SCF INFO PROCESS $NCP, LINESET command also displays the current file-system error. Example 20-4 shows an example of an SCF INFO PROCESS $NCP, LINESET display. Example 20-4.
PAGE 560
Troubleshooting SCF Commands Table 20-4. Common File-System Errors (page 2 of 2) Error Cause Recovery 140 This error indicates that the Expand line has been disconnected. The cause could be a problem with modem-tosystem communications or with a phone line, a cable, or an X.25 connection. Examine the physical connections and modem settings to ensure that cables are plugged into the correct line interface unit (LIU).
PAGE 561
Troubleshooting SCF Commands The SCF INFO PROCESS $NCP, NETMAP command is useful for displaying the current routing data. Example 20-5 shows a sample display of the SCF INFO PROCESS $NCP, NETMAP command. Asterisks indicate the best-path route. A plus sign (+) instead of an asterisk (*) would indicate that the Expand subsystem is attempting to connect or reconnect a path. Example 20-5.
PAGE 562
Troubleshooting Problem Check-List Summary The SCF PROBE PROCESS, $NCP command is useful for displaying the intermediate nodes in a path and the typical time to each destination node. Example 20-6 shows a sample display of the SCF PROBE PROCESS, $NCP command. Example 20-6.
PAGE 563
Troubleshooting Resolving Specific Network Problems Resolving Specific Network Problems This subsection provides checklists for solving the following specific network problems: • • • • • • • $NCP Problems Expand Line-Handler Process Problems SWAN Concentrator Problems WAN Subsystem Problems Expand-Over-X.25 Problems Expand-Over-IP Problems Multi-CPU Path Problems $NCP Problems Table 20-6 lists SCF commands that are useful for diagnosing problems with $NCP. Table 20-6.
PAGE 564
Troubleshooting Expand Line-Handler Process Problems Table 20-6. Identifying $NCP Problems With SCF Commands Command Use INFO PROCESS $NCP, SYSTEMS Displays all known systems. If no connection is established, the SYSTEMS option displays an infinite time factor and hop count. The SYSTEMS option is similar to the CONNECTS option, except that the CONNECTS option displays only the systems connected.
PAGE 565
Troubleshooting SWAN Concentrator Problems Table 20-7. Expand Line-Handler Process Problem-Resolution Procedures (page 2 of 2) Problem Procedure Expand Layer 2 protocol You can use the SCF STATS LINE command to examine Layer 2 statistics. Figure 20-1 (earlier in this section) is an example of an SCF STATS LINE command. SWAN Concentrator Problems This subsection provides ServerNet wide area network (SWAN) concentrator troubleshooting guidelines and identifies common SWAN concentrator problems.
PAGE 566
Troubleshooting SWAN Concentrator Problems Table 20-8. SWAN Concentrator Problem-Resolution Check List (page 2 of 2) Task Procedure Check the Ethernet paths configured for each CLIP on the SWAN concentrator. To determine the state of the Ethernet paths configured for each CLIP on a specific SWAN concentrator, use the following SCF commands: STATUS STATUS STATUS STATUS STATUS STATUS PATH PATH PATH PATH PATH PATH $ZZWAN.#conc-name.1.a $ZZWAN.#conc-name.1.b $ZZWAN.#conc-name.2.a $ZZWAN.#conc-name.2.
PAGE 567
Troubleshooting WAN Subsystem Problems Ethernet ports to the same segment, either omit ALTTCPIP or set it to a NonStop TCP/IP process that does not exist. Both solutions will result in the following EMS message: “Connected to Wrong ETHERNET PORT.” • • The SWAN concentrator’s Ethernet ports are reversed. If the ports are reversed, you will receive the following EMS message: “Connected to Wrong ETHERNET PORT.
PAGE 568
Troubleshooting WAN Subsystem Problems Table 20-9. WAN Subsystem Problem-Resolution Check List (page 2 of 2) Task Procedure Check the state of the default Subsystem Control Point (SCP) manager process ($ZNET). To determine if $ZNET is running, use the following command at the TACL prompt: STATUS $ZNET Typically, there should be a permanent SCP process called $ZNET on each Integrity NonStop server that SCF uses by default.
PAGE 569
Troubleshooting WAN Subsystem Problems Common WAN Subsystem Problems The following is a list of common WAN subsystem problems: • • The SNMPCODE, KERNELCODE, or PROGRAM file is not in the correct subvolume or is not secured for “N” read access (the leftmost character in the file security string). SNMPCODE and KERNELCODE are download files for the SWAN concentrator; PROGRAM is a microcode file where the data link control (DLC) task is located.
PAGE 570
Troubleshooting Expand-Over-X.25 Problems Expand-Over-X.25 Problems Table 20-10 provides general suggestions to help you solve problems with Expandover-X.25 lines. Table 20-10. Expand-Over-X.25 Problem-Resolution Procedures Task Procedure Check the X25AM process that controls the X.25 line to make sure that the line is operational and that the appropriate subdevice(s) are started. To verify that a subdevice is correct, use the X25AM subsystem SCF INFO SU command.
PAGE 571
Troubleshooting Expand-Over-IP Problems Expand-Over-IP Problems You can diagnose most Expand-over-IP line-handler process problems using information provided by the Expand subsystem SCF STATUS LINE command with the DETAIL option. This command provides error information in the Detailed State and Detailed Info fields. Example 20-7 shows an SCF STATUS LINE, DETAIL command display. (The Detailed State and Detailed Info fields are shown in boldface type.) Example 20-7.
PAGE 572
Troubleshooting Expand-Over-IP Problems Table 20-11. Detailed States (Expand-Over-IP) (page 2 of 3) Detailed State Cause/Effect Recovery CONNECTING The Expand-over-IP line-handler process has connected to the local NonStop TCP/IP process is and is now attempting to connect to the remote Expand-over-IP linehandler process. Use the SCF STATS LINE and SCF INFO LINE commands to further diagnose the problem. These commands are described in Resolving Expand-Over-IP Connection Problems on page 20-21.
PAGE 573
Troubleshooting Expand-Over-IP Problems Table 20-11. Detailed States (Expand-Over-IP) (page 3 of 3) Detailed State Cause/Effect Recovery QUERY A connection has been established with the remote Expand-over-IP line-handler process, but no data has been received within the inactivity interval. The Expandover-IP line-handler process is sending Probe messages to the remote Expand-over-IP linehandler process to verify that it is operational.
PAGE 574
Troubleshooting Expand-Over-IP Problems If the line is configured to issue active connect requests, determine if Connect Command frames (Conn Cmd column) are being sent (Send row). If the line is configured in passive connect mode, determine if Connect Command frames (Conn Cmd column) are being received (Rcvd row). If no Connect Command frames are being sent or received, the destination line-handler process may not be operational or there may be a network problem.
PAGE 575
Troubleshooting Expand-Over-IP Problems Table 20-12. Messages Displayed in the Detailed Info Field (Expand-Over-IP) (page 2 of 2) Message Description Ownership error The Expand-over-IP line-handler process is unable to switch processors.
PAGE 576
Troubleshooting Expand-Over-ATM Problems Refer to the Operator Messages Manual for cause, effect, and recovery information for the event numbers generated by the Expand subsystem. Expand-Over-ATM Problems You can diagnose most Expand-over-ATM line-handler process problems using information provided by the Expand subsystem SCF STATUS LINE command with the DETAIL option. This command provides error information in the Detailed State and Detailed Info fields.
PAGE 577
Troubleshooting Expand-Over-ATM Problems Table 20-14. Detailed States (Expand-Over-ATM) (page 2 of 3) Detailed State Cause/Effect Recovery CONNECTING The Expand-over-ATM line-handler process is attempting to connect to the remote (destination) Expandover-ATM line-handler process. Use the SCF STATS LINE and SCF INFO LINE commands to further diagnose the problem. These commands are described in Resolving Expand-Over-ATM Connection Problems on page 20-26.
PAGE 578
Troubleshooting Expand-Over-ATM Problems Table 20-14. Detailed States (Expand-Over-ATM) (page 3 of 3) Detailed State Cause/Effect Recovery QUERY A connection has been established with the remote Expand-over-ATM line-handler process, but no data has been received within the inactivity interval. The Expandover-ATM line-handler process is sending Probe messages to the remote Expand-over-ATM linehandler process to verify that it is operational.
PAGE 579
Troubleshooting Expand-Over-ATM Problems Cmd column) are being received (Rcvd row). If no Connect Command frames are being sent or received, the destination line-handler process may not be operational or there may be a network problem. If the Invalid Frames Rcvd counter is greater than 0, frames are being corrupted; contact your HP support representative.
PAGE 580
Troubleshooting Multi-CPU Path Problems Table 20-15. Messages Displayed in the Detailed Info Field (Expand-OverATM) (page 2 of 2) Message Description PVC unavailable, error nnn The ATM permanent virtual circuit (PVC) used by the Expand-over-ATM line-handler process is not available. Make sure the PVC is configured properly. The Expand-over-ATM line will become ready when the associated ATM line becomes available. ATM error codes (nnn) are described in the ATM Configuration and Management Manual.
PAGE 581
Troubleshooting Multi-CPU Path Problems Table 20-16. Multi-CPU Path Problem Resolution Procedures Task Procedure Check that the multi-CPU path is enabled. Use the following SCF command to see if the paths you expect to be part of the multi-CPU path are actually configured: INFO PROCESS $NCP, LINESET If some of the paths do not display an “S” next to the LINESET number, then either the local or remote Expand line-handler process does not have the SUPERPATH modifier enabled.
PAGE 582
Troubleshooting Reporting Network Problems balanced after a few minutes, then it is likely that most of the traffic is between just one or a few pairs of endpoints, making it impossible to spread the load over all the paths in the multi-CPU path. If this is not the case and the problem is severe, then stop and restart one path to force traffic to be redistributed.
PAGE 583
Troubleshooting Tracing Tracing $NCP To start a trace of $NCP, use the following command: TRACE PROCESS $NCP, TO $file_name, SELECT ALL, WRAP, RECSIZE 500 To stop the trace, use the following command: TRACE PROCESS $NCP, STOP $file_name specifies the name of the file to which the trace records will be written.
PAGE 584
Troubleshooting Resolving Common Network Problems Resolving Common Network Problems This subsection shows you, through examples, how to solve several common network problems. These problems include the following: • • • • Slow Response Time Network Congestion Node Not Available • Path Down • Line(s) Down Duplicate Node Slow Response Time Slow response time is indicated when network response is worse than expected relative to normal day-to-day performance.
PAGE 585
Troubleshooting Slow Response Time Step 2: Display routing data Once you have isolated the path to the nodes causing the bottleneck, use the following Expand subsystem SCF command to display the routing data of the network control process ($NCP) at one end of the slow path: INFO PROCESS $NCP, NETMAP, AT \system-name Example 20-5, SCF INFO PROCESS $NCP, NETMAP Display, on page 20-9 shows an example of an INFO PROCESS $NCP, NETMAP display.
PAGE 586
Troubleshooting Network Congestion Network Congestion Slow response time indicate network congestion, which can be caused by the following conditions: • • • • • • Too much traffic for the current network capacity. A node or nodes that are fully operational but unable to process the traffic, causing bottlenecks. Peak loading of the network, causing temporary congestion. A downed path causing rerouting of traffic to other nodes (see Path Down on page 20-35).
PAGE 587
Troubleshooting Node Not Available If the path displays a plus sign (+), you should first attempt to resolve the outstanding request problem by issuing an Expand subsystem SCF ABORT PATH command to the identified path and then issuing an Expand subsystem SCF START PATH command to start it again. If the path constantly displays the plus sign (+), this indicates that the connection cannot be established.
PAGE 588
Troubleshooting • • • Node Not Available NEXTSYS modifier value. A large number of Level-2 DISC (disconnect) supervisory frames usually indicates an incorrect NEXTSYS number. Interface selection (RS-232 or RS-422). The controller defines its pinouts by the proper setting of the interface. ASSOCIATEDEV modifier and the system load command files for Expand-overX.25, Expand-over-SNA, and Expand-over-IP, and Expand-over-ATM line-handler processes.
PAGE 589
Troubleshooting Adding Low-Speed Lines to a Multi-Line Path Adding Low-Speed Lines to a Multi-Line Path Adding more low-speed lines to a multi-line path can increase the number of OOS frames that a path needs in order to reassemble. The effect in this case is not on the buffer space used but on the total time taken for the sending Expand line-handler process to receive its ACK. As a result, the sending node may experience an increase in Layer 4 timeouts.
PAGE 590
Troubleshooting Duplicate Node Expand Configuration and Management Manual—529522-002 20-38
PAGE 591
A SCF Error Messages This appendix contains error messages returned by the SCF subsystem when you execute SCF commands. For other Expand network-related errors, refer to the Operator Messages Manual. Expand Error 00001 EXPAND 00001 Too many object names. Object Name: object-name. Cause. You specified more than 30 object names in an SCF command. Effect. The SCF command was not executed. Recovery. Re-enter the command using fewer that 30 object names. Expand Error 00002 EXPAND 00002 Negative LH response.
PAGE 592
SCF Error Messages Expand Error 00005 Expand Error 00005 EXPAND 00005 Object type and name mismatched. OBJNAME: object-name OBJTYPE: object-type. Cause. The subtype of an Expand line-handler process object name does not match the expected object type. For example, you may have entered a path logical device name after specifying a LINE object type in the SCF command. Effect. The SCF command was not executed. Recovery. Re-enter the command with matching object types and object names.
PAGE 593
SCF Error Messages Expand Error 00009 Expand Error 00009 EXPAND 00009 Negative $NCP response. OBJNAME: File system err: #R##. object-name. Cause. The SCF command was rejected by the network control process ($NCP). Effect. The SCF command was not executed. Recovery. Correct the file-system error, then re-enter the command. Expand Error 00010 EXPAND 00010 INTERNAL ERR: Rcvd Bad Network trace from SYSTEM #R# to SYSTEM #R#. Cause.
PAGE 594
SCF Error Messages Expand Error 00013 Expand Error 00013 EXPAND 00013 The SYSTEM system-number is not defined. Cause. The system number specified in the SCF command is not recognized by the network control process ($NCP). Effect. The SCF command was not executed. Recovery. Re-enter the command making sure to use the correct system number. Expand Error 00014 EXPAND 00014 All paths to the SYSTEM system-number are down. Cause. All paths to the specified system are down. Effect.
PAGE 595
SCF Error Messages Expand Error 00017 Expand Error 00017 EXPAND 00017 Not supported for a down-version system. Cause. The SCF command was rejected by the Expand manager process ($ZEXP). The information requested cannot be obtained from an older (down-version) system. Effect. The SCF command was not executed. Recovery. No action is required. Expand Error 00018 EXPAND E00018 Configuration error or Memory allocation failure. Cause.
PAGE 596
SCF Error Messages Expand Error 00021 Expand Error 00021 EXPAND E00021 System sysnum is not a multi-CPU path neighbor. Cause. The SCF INFO PROCESS $NCP, RPT command was rejected because there is no multi-CPU path connected to the specified system. Effect. The SCF command is not executed. Recovery. No action is required.
PAGE 597
B Expand and WAN SCF Comparison This appendix compares the commands provided by the SCF interface to the Expand subsystem with the commands provided by the SCF interface to the WAN subsystem. Refer to Section 14, Subsystem Control Facility (SCF) Commands for a general comparison of the two SCF interfaces. Command Comparison Table B-1 provides a command-to-command comparison of Expand SCF and WAN SCF commands and explains which command to use to perform a specific task. Table B-1.
PAGE 598
Expand and WAN SCF Comparison Command Comparison Table B-1. Expand and WAN SCF Command Comparison (page 2 of 6) Expand Subsystem SCF Command WAN Subsystem SCF Command Command Function ALTER ALTER Expand SCF: • • Use the ALTER LINE and ALTER PATH commands to make temporary changes to attributes and attribute values for a selected Expand line-handler process. Use the ALTER PROCESS command to make temporary changes to attributes and attribute values for the network control process ($NCP).
PAGE 599
Expand and WAN SCF Comparison Command Comparison Table B-1. Expand and WAN SCF Command Comparison (page 3 of 6) Expand Subsystem SCF Command WAN Subsystem SCF Command Command Function INFO INFO Expand SCF: • • • • Use the INFO LINE command to display current Layer 2 attributes and attribute values for a selected Expand line-handler process. Use the INFO PATH command to display current Layer 4 attributes and attribute values for a selected Expand line-handler process.
PAGE 600
Expand and WAN SCF Comparison Command Comparison Table B-1. Expand and WAN SCF Command Comparison (page 4 of 6) Expand Subsystem SCF Command WAN Subsystem SCF Command Command Function PRIMARY PRIMARY Expand SCF: • Use the PRIMARY PROCESS command to cause the backup processor to become the primary processor and the primary processor to become the backup processor for a selected Expand line-handler process or for the network control process ($NCP).
PAGE 601
Expand and WAN SCF Comparison Command Comparison Table B-1. Expand and WAN SCF Command Comparison (page 5 of 6) Expand Subsystem SCF Command WAN Subsystem SCF Command Command Function STATUS STATUS Expand SCF: • • • Use the STATUS LINE command to display the dynamic state, primary process ID (PPID), backup process ID (BPID), and other information about a selected Expand line.
PAGE 602
Expand and WAN SCF Comparison Command Comparison Table B-1. Expand and WAN SCF Command Comparison (page 6 of 6) Expand Subsystem SCF Command WAN Subsystem SCF Command Command Function VERSION VERSION Expand SCF: • Use the VERSION PROCESS command to display the version level of the Expand manager process ($ZEXP), the network control process ($NCP), or a selected Expand line-handler process. WAN SCF: • Use the VERSION SUBSYS command to display the version level of the WAN manager process ($ZZWAN).
PAGE 603
Expand and WAN SCF Comparison ALTER Command Comparison ALTER Command Comparison You can use the SCF interface to the WAN subsystem to permanently change the value of any Expand modifier used by an Expand line-handler process or the network control process ($NCP). Expand modifiers are described in Section 16, Expand Modifiers. Modifier-to-Attribute Comparison Most—but not all—Expand modifiers have corresponding attribute names in Expand SCF.
PAGE 604
Expand and WAN SCF Comparison Altering Timeout Periods Altering Timeout Periods Certain Expand SCF attributes are used to set a timeout period (for example, the OSTIMEOUT attribute specifies the Expand out-of-sequence packet timeout period). These Expand SCF attributes accept different units of time than the Expand modifiers with which they correspond.
PAGE 605
Glossary active connect request. The default connect request method used by an Expand linehandler process when it attempts to establish an end-to-end connection. When an Expand-over-NAM line-handler process issues an active connect request, the network access method (NAM) process attempts to initiate a connection.
PAGE 606
Glossary ATMSAP connection ATMSAP connection. A virtual circuit that is permanently established through the SLSA subsystem. ATMSAPs are the same type of connection as PVCs; they save bandwidth associated with circuit establishment and tear down in situations where certain virtual circuits must exist all the time. See also permanent virtual circuit (PVC) and switched virtual circuit (SVC). attribute. Parameters associated with an SCF object. Availability Statistics and Performance (ASAP).
PAGE 607
Glossary concentrator manager process (ConMgr) concentrator manager process (ConMgr). A process provided as part of the wide area network (WAN) subsystem. The ConMgr process runs in each processor that supports WAN products and provides management functions to the WAN subsystem and WAN products, such as downloading data link control (DLC) tasks to the communications line interface processors (CLIPs) on the ServerNet wide area network (SWAN) concentrator and selecting the preferred path for the DLC tasks.
PAGE 608
Glossary EMS. also the resources available on the path to accommodate more traffic. The ETF indicates the inverse proportion of traffic that should be sent over the path compared to an unloaded path with a TF of 1. Therefore, a path with a TF of 6 reports an ETF of 12 when it is half loaded. See also multipacket frame. EMS. See Event Management Service (EMS). End-to-End protocol.
PAGE 609
Glossary Expand line-handler process pair Expand line-handler process pair. The Expand line-handler processes at the source and destination node on a multi-CPU path. Expand line-handler processes at each source and destination node on a multi-CPU path are paired in order to guarantee message order; all messages between that source and destination node are sent through this Expand line-handler process pair.
PAGE 610
Glossary Expand priority Expand priority. The priority of an Expand message. The Expand priority is based on the priority level of the application process that created the message, unless the SETMODE 71 procedure is used. fabric. A complex set of interconnections through which there can be multiple and (to the user) unknown paths from point to point. The term fabric is used to refer to the X or Y portion of the ServerNet system area network (SAN); for example, the X-fabric.
PAGE 611
Glossary group group. The set of all objects accessible by a pair of service processors (SPs) located in the processor multifunction (PMF) customer-replaceable unit (CRU). There is one group in a system enclosure. Guardian. The original application program interface (API) to the NonStop Kernel operating system. HC. See hop count (HC). HDLC Extended Mode protocol. See High-Level Data Link Control (HDLC) Extended Mode protocol. HDLC Normal protocol. See High-Level Data Link Control (HDLC) Normal protocol.
PAGE 612
Glossary Input/Output adapter module (IOAM) Input/Output adapter module (IOAM). A product that allows the HP NonStop™ NS-series server to communicate with fiber-channel or Ethernet products. The IOAM is a rackmountable enclosure that contains two ServerNet Switch Boards and up to 5 FibreChannel ServerNet adapters (FCSAs) or Gigabit Ethernet 4-port ServerNet adapters (G4SAs). Up to 5 G4SAs, FCSAs, or any combination can be added per IOAM module, for a total of 10 adapters per IOAM enclosure.
PAGE 613
Glossary Layer 3 Layer 3. A term that is used to refer to the Network Layer of the Open Systems Interconnection (OSI) Reference Model. Layer 3 governs the switching and routing of information between systems in the network and is responsible for error checking and recovery. The Expand line-handler process’s Network Layer, or path functions, corresponds to the PATH object referred to by the Expand Subsystem Control Facility (SCF) interface. Layer 4.
PAGE 614
Glossary load factor load factor. The ratio between a path’s effective time factor (ETF) and its base time factor (TF). See also effective time factor (ETF). local system. A term used to refer to the system to which your terminal is directly connected. logical device name. The name assigned to an input-output process (IOP) during its configuration. logical device number. A number that identifies a particular input-output (I/O) device in the system. logical network partitioning.
PAGE 615
Glossary multi-CPU path multi-CPU path. The fundamental component of the Expand multi-CPU feature. A multiCPU path can consist of up to 16 individual Expand paths, including multi-line paths. Each Expand line-handler process (or multi-line path) that is a member of a multi-CPU path is configured in a different processor. See also Expand multi-CPU feature. multi-line path. A path between two neighbor systems that consists of more than one physical line.
PAGE 616
Glossary network control process (NCP) network control process (NCP). A process pair, named $NCP, that runs in each system of an Expand network.
PAGE 617
Glossary object object. (1) One or more of the devices, lines, processes, and files in an HP subsystem; any entity subject to independent reference or control by one or more subsystems. (2) In SCF, a resource controlled by an SCF subsystem. SCF objects include processes, disks, disk files, and data communications lines. Each object has an object type and an object name. See also object name and object type. object attribute. See attribute. object name.
PAGE 618
Glossary passthrough routing over-ATM line-handler process. See also active connect request and network access method (NAM). passthrough routing. A routing scheme used by the Expand subsystem that permits intermediate nodes to route, or passthrough, data packets to the destination system. This scheme reduces the number of lines required between systems because systems do not have to be directly connected. passthrough traffic. Packets received from a remote node that are destined for another remote node.
PAGE 619
Glossary process identification number (PIN) process identification number (PIN). A number that uniquely identifies a process running in a processor. The same number can exist in other processors in the same system. See also process ID. processor switch. A self-contained, rack-mountable enclosure that contains a ServerNet switch board and supporting power, cooling, maintenance, and interfacing components. profile. A disk file containing modifiers and default values.
PAGE 620
Glossary route route. The sequence of paths that data follows when traveling between source and destination nodes. There is only one active route at a time between communicating nodes. See also path and best-path route. RPT. See reverse pairing table (RPT). satellite-connect line-handler process. An Expand line-handler process that implements the satellite-efficient version of the High-Level Data Link Control (HDLC) protocol, HDLC Extended Mode.
PAGE 621
Glossary SH SH. See split horizon (SH). single-line path. A path that consists of one line. See also line, path, and route. Contrast multi-line path. SLSA subsystem. See ServerNet LAN systems access (SLSA) subsystem. SNA. See Systems Network Architecture (SNA). split horizon (SH). An alternative routing algorithm provided by the Expand subsystem. The main advantage of SH is that alternate paths are immediately known (temporary discontinuity never occurs).
PAGE 622
Glossary switched virtual circuit (SVC) switched virtual circuit (SVC). A virtual circuit that is dynamically established on demand. See also permanent virtual circuit (PVC). SYSTEM command. A network-related HP Tandem Advanced Command Language (TACL) command that designates a default system name. system load. (1) To start the system; to load the operating system image from disk or tape into the memory of a processor. (2) The process of loading a copy of the operating system into a halted processor.
PAGE 623
Glossary tuning supports a ServerNet wide area network (SWAN) concentrator. One TFTP server process is required for each NonStop TCP/IP process that supports a SWAN concentrator. tuning. A tactical adjustment of a network’s dynamic resources to achieve some welldefined performance goals. Tuning is influenced by and can influence the activities of network planning, configuration, management, and troubleshooting. UDP. See User Datagram Protocol (UDP). UDP port. See port number. User Datagram Protocol (UDP).
PAGE 624
Glossary wide area network (WAN) subsystem wide area network (WAN) subsystem. The subsystem that is used to configure and manage Expand line-handler processes and the network control process ($NCP). The WAN subsystem also provides access to the ServerNet wide area network (SWAN) concentrator. See also ServerNet wide area network (SWAN) concentrator. wide area network (WAN) subsystem manager process.
PAGE 625
Index A ABORT command 14-8/14-9 ABORT LINE command 18-26 ABORT PATH command 18-26 ABORTTIMER attribute ALTER PROCESS $NCP command and 14-21 INFO PROCESS $NCP command and 14-51 ABORTTIMER modifier 6-4, 17-27 ACK 17-14 ACTIVATE command 14-9 ACTIVATE PROCESS $NCP 18-27 Active connect request 17-50, 17-53, 17-59 ADD DEVICE command 18-19 ADDRESS attribute, INFO LINE command and 14-32 AFTERMAXRETRIES attribute ALTER LINE command and 14-14 INFO LINE command and 14-35, 14-37, 14-42, 14-46 AFTERMAXRETRIES_DOWN modif
PAGE 626
Index B B BCC Errors, STATS LINE command and 14-89 Best-path routing 2-7, 17-23, 20-34 Bind requests 17-50 Bottlenecks, avoiding 17-69 Buffer errors, STATS LINE command and 14-89 Buffer pool description of 17-46 EXTMEMSIZE 17-47 inadequate allocation of 20-12 insufficient space in 17-19 Buffers 17-46 Bulk transfers 19-5 Bus topology 3-14 C Calculating a path time factor, formula 17-21 CALLTYPE_ATMSAP modifier 16-5 CALLTYPE_PVC modifier 16-5 CALLTYPE_SVC modifier 16-5 Cancel request packet 17-14 CLBDWNLOA
PAGE 627
Index D Connection requests active 17-50, 17-53, 17-59 packets 17-12 passive 17-51, 17-54, 17-59 Connection reset packets 17-13 Connection response packets 17-12 CONNECTS option 14-48, 14-54 CONNECTTIME attribute ALTER PROCESS $NCP command and 14-21 INFO PROCESS $NCP command and 14-51 CONNECTTIME modifier 6-5 CONNECTTYPE attribute ALTER LINE command 14-14 INFO LINE command and 14-47 CONNECTTYPE_ACTIVEANDPASSIVE modifier 16-7, 17-50, 17-53, 17-59 CONNECTTYPE_PASSIVE modifier 16-7, 17-51, 17-54, 17-59 CONNE
PAGE 628
Index E DSRTIMER attribute ALTER LINE command and 14-16 INFO LINE command and 14-32 DV messages 17-26 E E4SA 7-3/7-4, 8-5, 10-3/10-4, 11-4 Effective time factor (ETF), displaying 18-13, 20-12 End-to-End protocol description of 17-12 resolving problems with 20-12 ENQ 17-14 Error Frames, STATS LINE command and 14-89 Error messages A-1/A-6 Errors BCC 20-33, 20-36 FCS 20-33, 20-36 file-system 20-7 ETF, displaying 18-13, 20-12 Ethernet 17-64, 20-14 Ethernet 4 ServerNet adapter (E4SA) 7-3/7-4, 8-5, 10-3/10-4,
PAGE 629
Index F Expand-over-X.25 line-handler process configuring 10-1/10-16 features of 3-2, 17-4 Expand-over-X.
PAGE 630
Index K INFO PROCESS command 18-14, 20-5, 20-11, 20-33 INFO PROCESS $NCP command 14-47/14-65, 18-12, 20-9 INFO PROFILE command 18-14 INFO SU command 20-18 Information frames (I-frames) 20-33 Information frames, STATS LINE command and 14-87 Interactive network access 2-1 INTERFACE attribute ALTER LINE command and 14-14 INFO LINE command and 14-32 INTERFACE_RS232 modifier 16-10 INTERFACE_RS422 modifier 16-10 Internet Protocol (IP) 17-52 IP network routes, redundancy in 8-4 IP networks 3-4 IPADDRESS attribut
PAGE 631
Index M L4TIMEOUT attribute ALTER PATH command and 14-11 INFO PATH command and 14-25, 14-27 L4TIMEOUT modifier 10-13, 11-16, 16-14 LABEL command 15-9 Latency 19-4, 19-6 Layer 1, Expand functions at 17-10 Layer 2 displaying frames at 14-87 Expand functions at 17-10 statistics, analyzing 20-33 windowing, effect on performance of 19-10 Layer 3, Expand functions at 17-9 Layer 4 Expand functions at 17-9 retries at 17-14 send window 17-20 statistics, analyzing 20-33 STATS PATH command and 14-73, 14-78 timeout f
PAGE 632
Index M MAXCONNECTS attribute ALTER PROCESS $NCP command and 14-21 INFO PROCESS $NCP command and 14-50/14-52 MAXCONNECTS modifier 6-6 MAXRECONNECTS attribute ALTER LINE command and 14-14 INFO LINE command and 14-34, 14-37, 14-42, 14-46 MAXRECONNECTS modifier 16-15/16-16, 17-51 MAXTIMEOUTS attribute ALTER PROCESS $NCP command and 14-21 INFO PROCESS $NCP command and 14-50/14-52 MAXTIMEOUTS modifier 6-6 Measure CPU entity of 19-26 features of 2-9, 20-10 LINE entity of 19-24 monitoring performance with 20-2 N
PAGE 633
Index N Multi-CPU paths automatically rebalancing 14-21 benefits and disadvantages of 3-9 configuring 17-72 congestion control and 16-25 definition 17-22 description of 2-8, 14-3, 17-2, 17-72 displaying paths in 14-48, 18-13 displaying rebalancing time for 14-53 displaying reverse pairing table (RPT) for 14-62 displaying status of 14-97 extended packet format and 16-25 guarantee message order 17-30 initiating an immediate rebalance of 14-9 load balancing for 17-72 parallel connections 8-4 performance tuni
PAGE 634
Index O NETWORKDIAMETER attribute ALTER PROCESS $NCP command and 14-21 INFO PROCESS $NCP command and 14-52 split horizon algorithm and 14-52 NETWORKDIAMETER modifier 6-6, 17-30 Network-related TACL commands REMOTEPASSWORD 18-7 SYSTEM 18-3, 18-6 WHO 18-4 NEXT command 15-9 NEXTSYS attribute ALTER PATH command and 14-11 INFO PATH command and 14-25/14-26 NEXTSYS modifier 16-16, 20-18, 20-36 NODE ACK 17-13 node name character limitation 18-2 definition 18-2 NODE STAT 17-13 Node status acknowledgment 17-13 pack
PAGE 635
Index P P Packet format 19-9 Packet size changing 19-28 Expand subsystem overhead and 19-8 Packets displaying traffic 14-93 incoming 14-92 per message 19-13 synchronization of 17-15 Packet-switched data networks 2-5, 3-2 Pair count balancing 17-31, 19-18/19-19 Parallel Library TCP/IP determining the preferred and alternate processes for WAN 1-8 overview 17-52 partially qualified file name 18-2 Passive connect requests 17-51, 17-54, 17-59 Passthrough traffic effect on performance of 19-20 handling of 17-40
PAGE 636
Index Q PEXPSSWN profile 5-3, 7-13 PEXQMATM profile 1-5 PEXQMIP profile 1-5 PEXQMNAM profile 1-5 PEXQMSAT profile 1-5 PEXQMSWN profile 1-5 PEXQSATM profile 1-4 PEXQSIP profile 1-4, 8-17 PEXQSNAM profile 1-4 PEXQSSAT profile 1-4 PEXQSSN profile 1-4 PEXQSSWN profile 1-4 Physical Layer, Expand functions at 17-10 PING 14-73, 14-78, 17-15, 17-24 Pool failures 20-12 PORT modifier 20-18 Port numbers 8-14/8-15, 17-53 PRIMARY PROCESS command 14-65, 18-26 Prioritization, message 17-39 Priority routing 2-8, 17-39 PR
PAGE 637
Index R QUALITYTIMER attribute ALTER LINE command and 14-14 INFO LINE command and 14-31, 14-37, 14-41 QUALITYTIMER modifier 16-20 R READBUFFERS attribute, INFO LINE command and 14-32 Rebalancing algorithm 3-10, 17-31 multi-CPU paths 18-27, 19-16 Reconfiguration network 18-19 online 2-10 RECORD command 15-11 Remote passwords 2-10, 17-43, 18-7/18-8 Remote processes 18-9 Remote programs, running 18-5 REMOTEPASSWORD command 2-10, 18-7, 18-10 Reply packets 17-43 Request packets 17-43 Resource use 19-1 Respons
PAGE 638
Index S ServerNet cluster coexistence with ATM or IP 4-3/4-6 configuration considerations 4-1 connected to external node 4-1/4-8 monitor process description 12-4 Layer 2 functions of 17-10, 17-48 topology examples 4-1/4-8 ServerNet LAN Systems Access (SLSA) subsystem 7-3, 9-1, 11-4 ServerNet Wide Area Network (SWAN) concentrator 7-5, 10-5 Session Layer, Expand functions at 17-9 SH algorithm 14-51, 17-29 Single-line path 17-21 SLSA subsystem 7-3, 11-4 SNA connections benefits and disadvantages of 3-3 multi
PAGE 639
Index T Subnetworks 18-10 Subsystem components, Expand 17-2 Subsystem Control Point (SCP) 20-16 Subtypes direct-connect line-handler process 7-10 Expand 18-15 Expand-over-ATM line-handler process 9-14 Expand-over-IP line-handler process 8-19 Expand-over-SNA line-handler process 11-13 Expand-over-X.
PAGE 640
Index U THRESHOLD attribute, INFO LINE command and 14-33 Throughput 19-1 Time factor (TF) displaying 14-59, 18-13 path 17-21 route 17-26 Time factors, order of selection 17-22 TIMEFACTOR attribute INFO LINE command and 14-30, 14-36, 14-40, 14-44 INFO PATH command and 14-27 Timeouts level 2 17-65 level 4 17-14 out-of-sequence (OOS) 17-17, 17-71 TIMERBIND attribute ALTER LINE command and 14-17 description of 17-50 INFO LINE command and 14-45 TIMERINACTIVITY attribute ALTER LINE command and 14-15, 14-17 desc
PAGE 641
Index V V V6DESTIPADDR attribute ALTER LINE command and 14-15 INFO LINE command and 14-39 V6DESTIPADDR modifier 16-27 V6DESTIPADDR parameter 1-11 V6SRCIPADDR attribute ALTER LINE command and 14-15 INFO LINE command and 14-39 V6SRCIPADDR modifier 16-27 V6SRCIPADDR parameter 1-11 Variable packet size 19-14 Variable packet size feature benefits of 3-12 configuring 17-66/17-68 PATHPACKETBYTES modifier for 7-13, 10-14, 11-16 Variable packet-size 19-5 VERSION command 14-111 VERSION PROCESS command 18-14 VERSION
PAGE 642
Index Special Characters Expand Configuration and Management Manual—529522-002 Index-18